Post on 29-Jul-2020
THE CHEETAHS OF THE NORTHERN TULI GAME RESERVE, BOTSWANA:
POPULATION ESTIMATES, MONITORING TECHNIQUES AND HUMAN-
PREDATOR CONFLICT
A thesis submitted in fulfilment of the requirements for the degree of
MASTER OF SCIENCE
of
RHODES UNIVERSITY
By
ELEANOR I. BRASSINE
December 2014
1
Dedicated to the memory of
Joseph Dudley Dexter
First cheetah captured by camera trap in The Northern Tuli Game Reserve, Botswana.
2
ABSTRACT
Remaining viable cheetah (Acinonyx jubatus) populations in Africa are threatened by direct
persecution through conflict with farmers and habitat degradation and fragmentation. Botswana
is considered a stronghold for free roaming cheetahs in Africa, yet the country has had relatively
limited research on its cheetahs, and information from the east of the country is lacking. Data on
the current status of populations is thus required to make informed management decisions. My
study provides estimates of population density, abundance, distribution and status for the
demographically open cheetah population of the Northern Tuli Game Reserve (NOTUGRE) in
Botswana. The effectiveness of two population monitoring methods, namely camera trapping and
a photographic survey, were also investigated. Moreover, I report on the level of conflict between
livestock farmers and predators on rural communal farmlands within and adjacent to NOTUGRE.
Data were collected between May 2012 and November 2013. Results indicate a low population
density of 0.61 ± 0.18 adult cheetahs per 100 km² and a minimum population size of 10
individuals (nine adults and one cub). Camera traps placed at cheetah scent-marking posts
increased detection rates and provided ideal set up locations. This approach, together with Spatial
Explicit Capture-Recapture (SECR) models, is recommended for future studies. The long-term
studies that are required to better understand the status of cheetahs in Botswana do not exist.
Thus, photographic surveys may provide an alternative method for providing baseline data on
population numbers, distribution and demography. The third aspect of my study gathered
information on levels of livestock loss and human tolerance of predators through the use of
interviews (n = 80). Conflict with subsistence farmers is a concern as livestock depredation is
relatively high (9.1% of total livestock owned) and farmers had an overall negative attitude
towards conservation of large predators. My results suggest that human-predator conflict in this
area is more complex than the direct financial loss from depredation. Hence, reducing depredation
rates alone is unlikely to change farmer tolerance of wildlife on farmlands. Improved, responsible
farm management, including self-responsibility for livestock rearing, and positive appreciation
for wildlife are necessary. The NOTUGRE cheetah population requires further research to
understand possible threats to the population. Furthermore, a better understanding of the
connectivity between cheetahs of NOTUGRE, South Africa and Zimbabwe is required. The
number of cheetahs within NOTUGRE is too small to sustain a viable population, hence
conserving cheetahs outside of the protected area should be a priority for the conservation of the
3
population. This can only be achieved through assistance and involvement from national
authorities, local people and conservation organisations.
4
ACKNOWLEDGEMENTS
As I embarked upon this project, it became readily apparent that its success would rely heavily
on the support of many people and organisations. I am truly grateful to all of you that believed in
me and decided my research was worth your support. In particular, I would like to acknowledge
and thank the following:
The National Geographic’s Big Cat Initiative for funding the project, I would like to particularly
thank Dr Luke Dollar for his continued assistance with this grant.
The Rufford Foundation’s small grant program for funding the project.
Zoofaris and Wildwonders, but more specifically, Jackie Navarro, director of Wildwonders, for
not only providing funds for this project but also promoting my project overseas to raise the
funds. Thank you for your support!
The financial assistance of the National Research Foundation (NRF) towards this research is
hereby acknowledged. Opinions expressed and conclusions arrived at, are those of the author and
are not necessarily to be attributed to the NRF.
Rhodes University Postgraduate Scholarship, with this I would like to particularly thank John
Gillam and his team for their assistance in securing my two bursaries.
Mashatu Game Reserve for providing accommodation, a research vehicle and administrative
help. I am particularly grateful for the use of a Land Cruiser, the best vehicle a researcher could
ask for. It never failed to surprise me where this vehicle could take me! It was a great privilege
to work and live at Mashatu. In particular, I would like to thank David Evans and Pete le Roux,
for presenting me with the opportunity to conduct research in The Northern Tuli Game Reserve
and granting me permission to reside and operate at Mashatu. Also Pete, for your help in starting
up the Northern Tuli Cheetah Project, and for always finding the time to answer to my emails.
I will always be thankful to the NOTUGRE land owners for granting me permission to conduct
research in the reserve and for the donation of cool drinks for participants during my
questionnaire survey.
5
Tuli Safari Lodge for generously sponsoring two cheetah satellite collars, I would like to
particularly thank Will Roughton and Nick Hilterman for the interest you showed in my work
and your continuous help in finding funds for the project.
Idea Wild for sponsoring camera traps and batteries.
The Wildlife and Reserve Management Research Group for supplying equipment including
camera traps and casings.
The Department of Wildlife and National Parks of Botswana and the Government of Botswana
for granting me permission to conduct research in Botswana.
Dr Dan Parker, my supervisor, not only for your academic guidance, but for your continued
encouragement and intellectual and moral guidance when I needed it most. Thank you for your
friendship and giving me unlimited intellectual freedom and for the support you showed when I
decided to take on this research. I would also like to thank you for getting my drafts back to me
in record speed!
Everyone who helped me in the field. I am most particularly grateful for your company as long
hours in the field alone can be monotonous and, at times, tedious. I have also learned that an
incredible wildlife sighting is not the same when you do not have someone to share it with. Mike
Dexter, you were by far my most valuable assistant and this project could not have been realised
without your continuous assistance and support. I would also like to thank the following people
who helped me with field work: Gwendoline Brassine for long hours in the Botswanan sun whilst
setting up camera traps; Tiro Maleke, Lauren Satterfield, Kaizer Sekanonyana, and Tim Forssman
for your assistance with my questionnaire survey.
Larry Patterson and Erik Verryen for providing veterinary work and for your expertise.
The photographic survey could not have happened without the contribution of photographs from
a number of people. Above all, I would like to thank all the professional photographers, not only
for contributing beautiful sharp images but also sharing with me your unedited images from your
archives.
6
The Mashatu Guides, in particular Fish Maila, Kaizer Sekanonyana, Bashi Patane, Elvis, Dan
Masupe and Richard Modeme for volunteering to help me with my research by either pointing
out cheetah scent marking posts or taking photographs of cheetahs whilst on game drives
Louise & Core Carelsen, Kate Barrow, Saskia von Sperber for providing me with company and
moral support when times were difficult.
Marco & Vivian Harms and Ian & Charlotte Pollard for your friendship and support, providing
me with unfailing hospitality
Adolf for servicing and repairing my research vehicle. You were always ready to help me out at
the drop of a hat!
Shem Compion and Zendré Lategan for your hospitality and for providing us lay over
accommodation and great company on numerous occasions.
Tico McNutt from The Botswana Predator Conservation Trust, for your assistance with some of
the survey designs and help in securing funding for the project.
A number of people provided valuable information on cheetah sightings; specifically I would like
to thank Stuart Quinn, Bruce Petty, Cliffy Phillips, Brian Rode and Chantelle Venter. Although
you all lived far away it was always a pleasure to see you after a long dusty (or muddy) drive.
Mike Lester, David Wyman, and Doug Devries for your help to locate Mapula, the radiocollared
cheetah that had been missing for a few months in an unnerving yet spectacular flight over the
reserve.
Martin Villet for reading through the drafts of this project and providing moral support.
Mathilde Brassine for your assistance with bursary applications and reading the draft of this
thesis.
Armand Kok and Jeff Jenness for your assistance with GIS whilst I was in Botswana.
Andy Royle for your assistance with SPACECAP.
7
Travis Perry for your help with statistical analyses
Galagos Wildlife-Conservation for the Elephant proof cement blocks, an ingenious idea to protect
the camera traps.
I am also grateful to Brenda & Peter Dexter for your support throughout the years, you have
always been there when we needed someone to fall back on. Thank you.
My love of the bush and wildlife could not have developed without my incredible upbringing
from my parents, Manu Brassine & Nicole Dehalleux. From an early age my sisters and I were
taken to some of the most remote places in Africa, experiencing the true African bush as we
would set off for extended holidays, I could not have asked for a better childhood! Through your
love and support I have been able to pursue my dreams.
Most of all I would like to thank Mike Dexter, for your continuous and persistent motivation
throughout this journey; you were my assistant in the field and later in the writing up. You
relocated to Grahamstown with me and provided the continuous support when I needed it most
and always acknowledged my success. Thank you for sharing with me some unforgettable
experiences in the bush.
Thank you to all my friends and work colleagues who helped proof reading my thesis and
discussing some of the findings of this project and to anyone I may have accidentally left out of
this list.
8
TABLE OF CONTENTS Abstract …………………………………………………………………………….……….…. 3
Acknowledgements ………………………………………………………………….……..….. 5
Table of Contents ……………………………………………………………………………… 9
Chapter 1. General Introduction
1.1 Cheetah global status ........................................................................................................... 13
1.2 Threats to cheetah populations ............................................................................................. 13
Human persecution ............................................................................................................. 13
Habitat loss .................................................................................................................... 14
1.3 Competition with large predators......................................................................................... 14
1.4 Conservation of cheetahs in Botswana ................................................................................. 15
1.5 Study rationale ..................................................................................................................... 16
Chapter 2. Study Area
2.1 Location ............................................................................................................................... 18
2.2 Land use ............................................................................................................................... 19
2.3 History ................................................................................................................................. 20
2.4 Climate................................................................................................................................. 21
2.5 Topography and Drainage .................................................................................................... 22
2.6 Geology and Soils ................................................................................................................ 24
2.7 Vegetation ............................................................................................................................ 25
2.8 Fauna .................................................................................................................................... 26
2.9 Study animal ........................................................................................................................ 27
Chapter 3. Trapping the Elusive Cat
3.1 Introduction .......................................................................................................................... 28
3.1.1 Camera trapping .................................................................................................... 29
3.1.2 Capture-recapture method ..................................................................................... 30
3.1.3 Objectives ............................................................................................................. 31
3.2 Methods ............................................................................................................................... 31
3.2.1 First Survey – Regular trap configuration ................................................................ 33
9
3.2.2 Second survey – non-random configuration using scent-marking posts ............... 35
3.2.3 Extended survey .................................................................................................... 37
3.2.4 Cheetah identification ........................................................................................... 38
3.2.5 Data analysis.............................................................................................................. 38
3.2.6 Buffer……… ........................................................................................................ 40
3.3 Results .................................................................................................................................. 42
3.3.1 First Survey – Regular trap configuration ................................................................ 42
3.3.2 Second survey – non-random configuration using scent-marking posts ............... 42
3.3.3 Extended survey .................................................................................................... 43
3.3.4 SECR analysis for the second camera trapping survey ......................................... 44
3.3.5 Model refinement .................................................................................................. 45
3.3.6 SECR analyses for the extended survey ............................................................... 46
3.4 Discussion ............................................................................................................................ 46
3.5 Conclusion................................................................................................................................. 50
Chapter 4. Citizen Science in Cheetah Research
4.1 Introduction .......................................................................................................................... 52
4.2 Methods ............................................................................................................................... 52
4.2.1 Data collection ........................................................................................................... 54
4.2.2 Data collation ............................................................................................................ 54
4.2.3 Individual identification ........................................................................................ 55
4.2.4 Long term trends and population demography ..................................................... 56
4.2.5 Photographic survey - Minimum population size (NOTUGRE) ........................... 57
4.2.6 Distribution and home range................................................................................. 58
4.3 Results .................................................................................................................................. 59
4.3.1 Long term trends and population demography ..................................................... 59
4.3.2 Minimum population size (NOTUGRE) ............................................................... 61
4.3.3 Distribution and home range................................................................................. 62
4.3.4 Home range overlap .............................................................................................. 64
4.3.5 Cross boundary movement ................................................................................... 65
10
4.4 Discussion .......................................................................................................................... 65
4.4.1 Camera traps as an additional tool ..................................................................... 66
4.4.2 Minimum population size and trends ................................................................. 66
4.4.3 Family relations .................................................................................................. 67
4.4.4 Estimated age ......................................................................................................... 67
4.4.5 Distribution and home ranges ............................................................................ 67
4.4.6 Home range overlap ........................................................................................... 68
4.4.7 Sightings outside of NOTUGRE ........................................................................... 69
4.5 Conclusion ............................................................................................................................... 70
Chapter 5. Human-Predator Conflict
5.1 Introduction ......................................................................................................................... 72
5.2 Methods .............................................................................................................................. 72
5.2.1 Study area ........................................................................................................... 76
5.2.2 Data collection ....................................................................................................... 78
5.2.3 Conflict of responses .......................................................................................... 81
5.2.4 Data analysis .......................................................................................................... 81
5.2.6 AIC analyses - attitude index ............................................................................. 82
Husbandry index ..................................................................................................................... 83
Knowledge index .................................................................................................................... 83
5.3 Results ............................................................................................................................... 83
5.3.1 Demographics of respondents ............................................................................ 84
5.3.2 Livestock husbandry .......................................................................................... 84
5.3.3 Kraal design ........................................................................................................... 86
5.3.4 Wildlife and predators ....................................................................................... 88
5.3.5 Livestock predation ............................................................................................ 90
5.3.6 Circumstances of predation ................................................................................... 92
5.3.7 Trends in perceived predation levels .................................................................. 93
5.3.8 Predators removed by farmers ............................................................................ 93
5.3.9 Livestock loss .................................................................................................... 93
11
5.3.10 Knowledge of local predators ............................................................................ 94
5.3.11 Attitudes and perception towards wildlife ............................................................ 94
5.3.12 Attitude, knowledge and husbandry Indices ...................................................... 95
5.4 Discussion ........................................................................................................................... 97
5.4.1 Livestock husbandry practices ............................................................................. 99
5.4.2 Wildlife and predators.......................................................................................... 99
5.4.3 Level of conflict ................................................................................................... 100
5.4.5 Trends ………….. ............................................................................................... 103
5.4.6 Farm management recommendations ...................................................................... 103
5.4.7 Management implications ........................................................................................ 104
5.4.8 Aspects of attitude, knowledge and husbandry .................................................... 104
5.4.9 Compensation implications ..................................................................................... 105
5.4.10 Outreach…. ........................................................................................................ 107
5.5 Conclusion ............................................................................................................................... 108
Chapter 6. Synthesis and Conclusion
6.1 Camera trapping: a monitoring technique for cheetahs ...................................................... 110
6.2 Citizen science in cheetah research ........................................................................................ 111
6.3 Conservation status of cheetahs in NOTUGRE .................................................................... 113
6.4 Conclusion............................................................................................................................... 115
References ………………………………………………………………………………….. 118
Appendices ……………………………………………………………………………….... 133
12
CHAPTER 1 GENERAL INTRODUCTION
1.1 Cheetah global status
The cheetah (Acinonyx jubatus) is currently listed as vulnerable by the International Union for
the Conservation of Nature (IUCN), Global Red List (IUCN 2013) and figures as an Appendix I
species on the Convention on International Trade in Endangered Species of Fauna and Flora
(CITES). In addition, available data suggests that cheetahs are close to being classified as
endangered and Ray et al. (2005) identify the cheetah as a species in crisis with a high overall
level of priority based on its high vulnerability and exposure to a suite of external threats. The
cheetah used to be widely distributed across Africa and south west Asia, however in the past few
decades the species distribution range and total numbers have reduced dramatically from
approximately 15 000 in the early 1970s (Myers 1975) to a maximum of 12 000 by 1998 (Marker
1998) and more recently (2008) to approximately 7500 individuals (Buk & Marnewick 2010).
However, from a species conservation perspective, it is important to note that it is the remaining
viable sub-populations that should be considered rather than just individual animals (Marker
1998). Viable cheetah populations are only found in sub-Saharan Africa with Tanzania, Kenya,
Namibia and Botswana considered strongholds (Marker 1998).
1.2 Threats to cheetah populations
The critical status of the cheetah is based on two main components. Firstly, the cheetah is highly
vulnerable and this is mostly due to extensive range loss (>75% of their range in the last 150
years), a relatively high degree of specialisation and low reproductive rates (Caro 1994; Ray et
al. 2005). Secondly, cheetahs are exposed to a high number of threats which are contributing to
the species’ global decline (Ray et al. 2005). The main threats identified are habitat loss, conflict
with livestock farmers and competition with other large predators (Myers 1975; Ray et al. 2005).
Human persecution
The greatest threat to cheetahs is arguably their direct persecution by humans either from
exporting of live animals or killing (Woodroffe & Ginsberg 1998; Ray et al. 2005; Marnewick
et al. 2007; Purchase et al. 2007; Marker et al. 2010). Cheetahs are often eliminated on livestock
farms as they are perceived to pose a threat to livestock (Marker et al. 2003a; Ray et al. 2005;
13
Holmern et al. 2007). Game farmers, who rely on live sales of antelope or trophy hunting, will
also often consider the cheetah a liability (Marker et al. 2003a; Marnewick et al. 2007). Cheetahs
are eliminated mostly by shooting on sight, but also by vehicle collisions, trapping with cages
and then shooting trapped animals, snaring and poisoning (Ray et al. 2005; Marnewick et al.
2007; Marker et al. 2010). The wild cheetah population also suffers from over-exploitation, with
cheetahs being captured and exported both legally and illegally to international captive facilities
(Klein 2007; Marnewick et al. 2007; Purchase et al. 2007).
Habitat loss
Cheetah populations suffer from the loss of natural habitat and habitat fragmentation resulting
from human encroachment and development linked to an increasing human population (Meyers
1975; Woodroffe & Ginsberg 1999; Marnewick et al. 2007; Macdonald et al. 2010a).
Furthermore, an increasing human population is often accompanied by the depletion of the prey
base through over-exploitation or habitat destruction (Kelly & Durant 2000; Broomhall et al.
2003; Ray et al. 2005; Marker et al. 2010). Human encroachment can also alter the habitat in
terms of bush encroachment and the introduction of artificial waterholes which can cause a
change in prey species composition and may make the habitat less suitable for cheetahs (Buk &
Marnewick 2010). For example, accessibility to water causes an increase in antelope densities
and consequently higher densities of larger predators such as lions (Panthera leo) and spotted
hyenas (Crocuta crocuta) (Buk & Marnewick 2010). Cheetahs prefer areas of relatively low prey
density that are normally avoided by other large predators (Durant 1998), a predator avoidance
strategy to reduce kleptoparasitism and cub mortality (Durant 1998), so these changes in habitat
may be detrimental.
Competition with large predators
The viability of a cheetah population is also affected by the density of other large predators (Caro
1994). Cheetahs are subjected to a high rate of intra-guild competition and kleptoparasitism from
larger carnivores such as lions and spotted hyenas (Marnewick et al. 2007; Houser et al. 2009;
Durant et al. 2010). Larger predators also contribute to cheetah cub mortality and in some
instances adult mortality (Caro 1994; Laurenson 1994; Laurenson et al. 1995; Kelly & Durant
2000; Ray et al. 2005; Macdonald et al. 2010a), hence cheetahs are often more successful outside
protected areas where other large predators have been extirpated or occur at lower population
densities (Laurenson et al. 1995; Marker 1998; Marnewick & Cilliers 2006). Outside of reserves,
however, cheetahs frequently come into contact with livestock farmers, who may consider them
a threat to their livelihoods leading to their persecution as mentioned previously (Marker 1998;
Woodroffe & Ginsberg 1998; Selebatso et al. 2008). The amount of potential habitat for cheetahs, 14
including protected and un-protected areas is, therefore, influenced by, but not limited to, the
population status of other large predators and the level of human activity.
1.3 Conservation of cheetahs in Botswana
Although there have been detailed cheetah population studies conducted in East Africa (Caro
1994; Gros 1996, 2000, 2002; Kelly et al. 1998; Kelly & Durant 2000) and Namibia (Marker et
al. 2003b, 2007, 2008a, 2008b), significant knowledge gaps still remain for many parts of their
range (Ray et al. 2005; DWNP 2009). Botswana has had relatively limited research on its cheetah
populations, and information specific to the east of the country is sorely lacking. However,
Botswana is believed to be a key country for the remaining viable populations of cheetahs,
holding the second largest population of cheetahs in southern Africa after Namibia (Marker 1998;
Purchase et al. 2007; Winterbach et al. 2014), with a national population estimate of 1768
cheetahs (Klein 2007). Furthermore, the Botswana cheetah population is believed to be a large
contiguous population with the cheetah populations of South Africa, Namibia, Zambia and
Zimbabwe (DWNP 2009).
Botswana has designated approximately 17% of its land to wildlife protection (Game Reserves
and National Parks) and an additional 21% designated as Wildlife Management Areas (WMA)
where sustainable wildlife use is permitted (Klein 2007). However, cheetahs are found
throughout the country, with about half of the cheetah population occurring outside of formally
protected areas (Myers 1975; Marker 1998; Winterbach et al. 2014). Therefore, agricultural
zones are important areas for cheetahs and conservation efforts should encourage the co-existence
of cheetahs and humans (Caro 1994; Winterbach et al. 2014). Livestock farming in Botswana
has grown exponentially in the last few decades along with the accompanied change in land use
from previously unoccupied wildlife areas to livestock farmlands (Klein 2007). The national
livestock herd is estimated at 4.7 million (DWNP 2012) yet the maximum sustainable herd was
evaluated at 3.3 million cattle (Bus taurus) (World Bank 1983). Lack of livestock management
in Botswana has resulted in deterioration of the veld (open landscape covered in grass or low
scrub) and habitat degradation is evident by the decline in perennial grasses, lowered water tables,
widespread thorn bush (Acacia spp.) encroachment and an overall decrease of wildlife (Myers
1975; Klein 2007). Veterinary cordon fences, to control the spread of foot and mouth disease,
have further aggravated the situation as these barriers can prevent the natural movement of
wildlife (Bartlam-Brooks et al. 2011; Cozzi et al. 2013). Additionally, the increase in the human
population and pastoral activities in previously uninhabited wildlife areas has led to an increase
in human-predator conflict particularly as a result of an increase in encounter rates between 15
livestock and predators (Klein 2007). Human-predator conflicts can have severe negative
consequences to large predator populations due to direct persecution by farmers (Ogada et al.
2003; Thorn et al. 2014).
The Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES)
has granted Botswana five cheetahs for annual export as live specimens and hunting trophies.
Despite this cheetahs are legally protected under Botswanan legislation and may not be killed
under any circumstances, however the species suffers from both illegal poaching and persecution
by livestock farmers (Klein 2007). The protection of remaining viable populations of cheetahs
requires their conservation outside protected areas, particularly populations which straddle
international boundaries and experience different acting laws and persecutions (DWNP 2009).
1.4 Study rationale
Despite the importance to conservation and management planning, the status of the cheetah in
Botswana is poorly researched. Klein (2007) provides a summary of the past cheetah research
undertaken in Botswana: Population censuses have been carried out in the Central Kalahari Game
Reserve (CKGR), the Kgalagadi Transfrontier Park (KGTF), Ramsar Site in the Okavango Delta,
and Jwana Game Park through the use of spoor surveys (Klein 2007). More recently, follow up
studies have been conducted in CKGR (2012), KGTP (2013) and areas around the CKGR (2014)
(R. Klein, Cheetah Conservation Botswana, pers. comm.). Moreover, a camera trapping study is
currently being carried out in the Ghanzi farmlands area (R. Klein pers. comm.). Information on
the status and distribution largely comes from interviews, opportunistic sightings, and Problem
Animal Control (PAC) reports. Of particular concern is that the focus of cheetah research has
only been carried out in certain areas of the North, Central and South of the country (Figure 1.1
taken from Klein 2007) and information on the status of cheetahs, including estimates of total
cheetah numbers in Botswana, is derived from two spoor surveys undertaken in the CKGR and
KGTF (Klein 2007) (Figure 1.1). Furthermore, monitoring programs to determine population
trends have yet to be conducted, with the only information on population trends obtained from a
status questionnaire survey conducted by Cheetah Conservation Botswana (CCB) in 2006 (Klein
2007). Although information from such a survey can provide quick and useful baseline
information, it does not replace the need to establish adequate monitoring programs to understand
trends in cheetah populations and possible threats to these populations (Gros et al. 1996).
16
Figure 1.1 Map of Botswana showing regions (in green) which have had more focused cheetah research and derived density estimates (Reproduced from Klein 2007).
Population size and trends in population sizes are recognized as the most important predictors of
species extinction risk (O’Grady et al. 2004), yet there is clearly a gap in our knowledge of the
population size and status of Botswana’s cheetahs, particularly in the east of the country where
research has mostly been absent (Figure 1.1). Thus, the cheetah population of Botswana requires
more in-depth information on population sizes and distribution in different habitat types and land
use areas (Klein 2007; DWNP 2009). In addition, assessments of the impact of predator-conflict
on cheetah populations in communal farmlands are urgently needed (Klein 2007).
In this study, I provide information on the status of cheetahs in the most eastern region of
Botswana. Information on the occurrence, distribution, population size and density estimates, and
apparent trends in numbers are documented. I also report on population demographics and, where
feasible, estimated age and family relations of specific individuals. My study also seeks to
develop an effective monitoring tool for cheetahs by addressing the efficiency of various field
methods and sampling designs to effectively monitor cheetah populations. Specifically, I evaluate
the suitability and effectiveness of camera trapping surveys and photographic surveys for
providing quick and reliable estimates on cheetah population status, size and density. Finally, my
study documents human-predator conflict within the livestock farming communities bordering
the Northern Tuli Game Reserve (NOTUGRE) in Botswana.
17
CHAPTER 2 STUDY AREA
2.1 Location
The study was undertaken in the Northern Tuli Game Reserve (from here on referred to as
NOTUGRE), a private game reserve situated in the eastern corner of Botswana. The region lies
between latitudes 21º55’ and 22º15’S, and longitudes 28º 55’and 29º15’E (Figure 2.1) and forms
the eastern limit of The Tuli Block, a 350 km strip of privately owned land located north of the
Limpopo River (McKenzie 1990).
Figure 2.1 The location of the Northern Tuli Game Reserve (NOTUGRE) in eastern Botswana. (ArcGIS 10; map units: decimal degree; not projected).
NOTUGRE is naturally delineated by the Shashe River in the east and the Limpopo River in the
south (Figure 2.2). The former forms the border between Botswana and Zimbabwe and the latter,
the border between Botswana and South Africa. The South African border is fenced (total fence
length ~86km) but poorly maintained and does not restrict animal movement (Jackson et al.
2012). The northern boundary consists of a cut-line demarcating the Tuli Circle Safaris Area in
Zimbabwe. Animals move freely across this boundary and there is limited human activity as the
area is only used seasonally for trophy hunting purposes. On the western boundary there is an
electrified game fence (height: 2.1m; 3 electrical stands at 1.8m, 50cm, 20cm) intended to prevent
wildlife movement out of the reserve as well as livestock into the reserve. However, it is
frequently damaged by elephants (Loxodonta africana) and other wildlife and therefore does not
normally restrict the movements of large carnivores and/or livestock. The south-western and
18
eastern boundaries are unfenced. The study area also has a double veterinary cordon fence which
runs north to south in the west of the reserve (Figure 2.2). This fence was built to control foot
and mouth disease by preventing large herbivore movements (Kgathi et al. 2012). However,
small ungulates and some large ungulates, such as kudu (Tragelaphus strepsiceros) and eland
(Tragelaphus oryx) are able to cross this fence (pers. obs.). The game fences of NOTUGRE also
do not restrict movement of large carnivores; cheetahs (Acinonyx jubatus), lions (Panthera leo),
leopards (Panthera pardus), spotted hyenas (Crocuta crocuta) and African wild dogs (Lycaon
pictus) frequently move across these fences (Jackson et al. 2012).
Figure 2.2 Map of NOTUGRE (green) with the major drainage channels. A game fence runs along the western boundary (single dashed line) and a poorly maintained fence runs along the southern banks of the Limpopo River. There are no fences along the Zimbabwe boundaries. A double veterinary cordon fence (double dashed line) runs north to south in the west of NOTUGRE. Non-member properties (Portion 2 of Lowensa-la-Moridi and Talana Farm) are shown in orange. (ArcGIS 10; map units: decimal degree; not projected).
2.2 Land use
NOTUGRE was established by multiple landowners in 1986 (Steyn 2004). It consists of 36
individual properties and encompasses an area of 728km². These individual properties are used
for commercial ecotourism or private holiday purposes (Steyn 2004). The aim of the reserve is
the conservation of wildlife. There is no farming or pastoral activity in the reserve (Steyn 2004),
although one of the properties has some livestock and a citrus orchard (Fairfield property) which
are enclosed within a game fence. Within the NOTUGRE boundary, there are two non-member
properties (Figure 2.2); Portion 2 of Lowensa-la-Moridi and Talana Farm. The former is situated
north of the Limpopo River and west of the veterinary cordon fence. A small village, Lentswe
Le Moriti, is located in the south eastern corner of the property and the land is also used for crop 19
and pastoral farming. Talana farm is an agricultural farm situated west of the Motloutse River
and along the banks of the Limpopo River. This property is surrounded with an electrified game
fence.
Adjacent communal farmlands north west of NOTUGRE also formed part of the study area. The
main form of land use on the communal farms is subsistence agricultural and livestock
pastoralism; livestock kept includes goats (Capra hircus), sheep (Ovis aries), cattle (Bus taurus)
donkeys (Equus asinus) and chickens (Gallus gallus domesticus). Subsistence agricultural and
livestock farming also occurs east of the reserve across the Shashe River in Zimbabwe. South of
the Limpopo River are privately owned South African farms which are used for commercial crop
farming, sport hunting, and game and livestock farming. All farms in South Africa are relatively
well fenced. Mapungubwe National Park in South Africa also borders onto the Limpopo River
to the south of NOTUGRE.
2.3 History
The area has been occupied by human settlements since before 800AD, first with small groups
of Stone Age people and then later by Iron Age people (McKenzie 1990). There is also evidence
of settlements of the Babirwa Bantu people, who farmed and kept large numbers of livestock
until 1926 (McKenzie 1990). Archaeological artefacts found on the reserve suggest intensive
pastoralism from between 900AD and the 1800s (Dr T. Forssman, archaeologist, pers. comm.).
In the late 1880s and early 1890s, there was large scale movement of European settlers to the
area with periodic attempts at cattle farming (Lind 1974). With this influx of settlers came the
construction of artificial waterholes, roads, human habitations and an overall increase in human
activity (Lind 1974). Large predators (species unspecified) were heavily persecuted during this
time, with the highest persecution occurring in areas with higher human activities (Lind 1974).
In the 1950s, at least a 150 lions were shot (Lind 1974). In the mid-1960s, farming and most
hunting ceased when a number of landowners campaigned for the development of a wildlife
sanctuary (Lind 1974). By this stage, most of the fauna of the region had been decimated and
large predators were almost extinct (Lind 1974). In 1975, the Northern Tuli Conservation
Association was formed with the purpose of conserving wildlife and all hunting ceased by 1987
(McKenzie 1990).
The mammalian fauna have largely recovered but some species did not return to the area,
including the white rhinoceros (Ceratotherium simum), African wild dog, roan antelope
(Hippotragus equinus), sable antelope (Hippotragus niger) and giraffe (Giraffa camelopardalis). 20
Giraffes and African wild dogs were reintroduced in NOTUGRE in 1984 and 2008, respectively.
However, the wild dog population suffered severehuman persecution and as of 2012, no resident
wild dog pack remained within NOTUGRE.
2.4 Climate
The climate of NOTUGRE can be described as semi-arid and sub-tropical with temperature
fluctuating between -5ºC and 42ºC (McKenzie 1990). Temperatures peak during December and
January, and reach their minimum during June, July and August (Figure 2.3). Occasional light
frosts have been recorded during harsh winters (Lind 1974). Rainfall is low and unpredictable
and the majority falls in the summer months between November and March, usually induced by
convectional movements (Figure 2.4). The average annual precipitation is 386.5mm (for the years
1996-2013) with peak rainfall years receiving as much as 917mm (2000) and low rainfall/drought
years as little as 172mm (2012) (Figure 2.5). Light showers may occur occasionally in April and
September. Rainfall occurs mostly as afternoon thunderstorms with localised showers. However,
widespread soaking rains may occasionally occur, but these are uncommon and normally only
occur in wetter years (McKenzie 1990). The prevailing wind is south-easterly, with whirlwinds
common in the dry winter months (Lind 1974). Data for the below figures (Figure 2.3 – 2.5) were
taken from four weather stations within NOTUGRE. Weather stations with incomplete records
were removed from the dataset and only used where appropriate.
Figure 2.3 The mean monthly maximum and minimum temperatures (°C) for NOTUGRE over a 17 year period (1996-2013) taken from two weather stations (Mashatu Main Camp; Mashatu Tent Camp) within NOTUGRE.
21
Figure 2.4 Average monthly rainfall for NOTUGRE taken from three weather stations (Mashatu Main Camp; Mashatu Tent Camp; Jwala Lodge) over a 17 year period (1996 – 2013).
Figure 2.5 Total annual rainfall (mm) for NOTUGRE averaged from four weather stations (Mashatu Main Camp; Mashatu Tent Camp; Jwala Lodge; Limpopo Valley Airfield) from 1996 to 2013. Annual rainfall is calculated over each rainy season rather than the calendar year.
2.5 Topography and Drainage
The study area is bisected by a number of river channels with the run-off of the entire area
draining into the Limpopo River (McKenzie 1990) (Figure 2.6). The Shashe and Motloutse rivers
are two of the Limpopo River’s largest tributaries within the study area (Figure 2.2). The Majale
and Pitsane rivers are also major rivers that drain south-easterly into the Limpopo River. Other
minor rivers flow directly into the Majale, Motloutse, Limpopo and Shashe rivers. These rivers
run in a general north–south direction. All are non-perennial, flowing only sporadically for few
hours or days following rainfall during the summer months (McKenzie 1990). During winter,
rivers are dry with the exception of isolated pools in some of the major rivers (McKenzie 1990).
Artificial waterholes and natural waterholes (n = 56) are also scattered throughout the reserve, 22
although many pump fed waterholes were discontinued following the formation of NOTUGRE
(P. Le Roux, Mashatu General Manager, pers. comm.).
Figure 2.6 Map of the Northern Tuli Game Reserve showing the river channels, main broad vegetation types and waterholes on the reserve (ArcGIS 10; map units: decimal degree; not projected)
NOTUGRE has an average elevation of about 600m.a.s.l. The topography is predominantly flat,
particularly around the major rivers. This flat landscape slows the drainage and causes silt
deposition and the formation of marshes, locally known as vleis. Two prominent vleis hold water
during the summer months; one is situated near the Majale-Limpopo River confluence (Figure
2.7) and the other is found east of the Motloutse River.
Figure 2.7 The vlei near the Majale-Limpopo River. Elephant grass, Sporobolus consimilis, forms the main vegetation type. Photo: Mike Dexter
23
2.6 Geology and Soils
The geological formations of the study area belong to the Basement Complex and Karoo
Supergroup, with the basement exposures belonging to the Central Zone or Messina Group of the
Limpopo Mobile Belt, with several metamorphosed sedimentary rock types (Joubert 1984). The
geology comprises deep Clarens sandstone formations overlain by Letaba and Sabi River basalt
formations, cut by a number of east-west dolorite dykes (Joubert 1984). The landscape is
relatively flat apart for protrusions of the more resistant dykes that form long narrow ridges
intersected by the main river channels (Joubert 1984). Sandstone outcrops, locally known as
koppies, also occur along the Limpopo and Motloutse rivers (Figure 2.8) (Joubert 1984). Along
all the major river systems are alluvial floodplains which typically have nutrient rich soils
(Joubert 1984).
Figure 2.8 A view near the Limpopo-Motloutse confluence showing a sandveld valley and a sandstone koppie with Acacia species as the dominant species (foreground). In the background spreads a wide expanse of Mopane Veld with Colophospermum mopane as the main woody vegetation. Photo: Eleanor Brassine
Over-grazing of the herbaceous layer by both wild herbivores and livestock has resulted in
accelerated and extensive erosion which is particularly evident in the form of sheet and donga
erosion and large bare areas completely devoid of vegetation (Figure 2.9) (McKenzie 1990).
Basalt areas have a very thin layer of top soils remaining, with Glenrosa, Mispah and Mayo
forming the dominant forms, while riverine soils are deep and belong mainly to the Oakleaf,
Valsriver and Rensburg forms (Joubert 1984; Mckenzie 1990). Alexander (1984) classified the
soil into nine descriptive classes which were subsequently regrouped into three major soil types.
These are residual soils, alluvial soils, and eluvial soils.
24
Figure 2.9 Example of a large bare area completely devoid of vegetation, apart from small clumps of the common annual forb, Dubbeltjie (Tribulus terrestris), as seen in the foreground. Photo: Mike Dexter
2.7 Vegetation
NOTUGRE falls within the Mopane Bioregion of the Savannah biome, classified as arid, base
rich savannah (Gotze et al. 2008). The vegetation can be broadly classified as Mopane Veld, but
is also made up of a wide variety of other smaller habitats (McKenzie 1990). The main rivers are
flanked by riverine forests forming a thick canopy (Figure 2.10). Species in this habitat include
Mashatu trees (Xanthoceris zambeziaca), groves of Fever trees (Acacia xanthophloea), and Mlala
palms (Hyphaene banguelensis) (Seliers 2007). Croton (Croton megalobotrys) thickets can also
be found along the banks of the more prominent water channels including the Majale and Pitsane
rivers (Figure 2.10; pers. obs.).
Figure 2.10 Alluvial plains with Boscia foetida savannah (foreground) and Croton megalabotrys thicket (background) along the banks of the Majale River. Photo: Eleanor Brassine
25
Fifteen different vegetation types were described by Alexander (1984) based on the species
present and their relative abundance, but it is likely that changes in the vegetation have occurred
since that study which was of a preliminary nature. Furthermore, distinguishing between the
different habitat types can be difficult as they often merge into one another. McKenzie (1990)
later regrouped Alexander’s 15 vegetation types into either predominantly open or closed.
Predominantly open habitats: Boscia foetida savannah, Colophospermum mopane Terminalia
prunoides middleslopes, Colophospermum mopane scrubveld, Salvadora angustifolia bushveld.
Predominantly closed habitats: Valley bush, Acacia tortillis savannah, and Croton megalabotrys
thicket. Joubert (1984) also groups these habitats into three landscapes, namely: Floodplain on
alluvium; Colophospermum mopane/Terminalia prunioides Rugged Veld on Basalt; and Karoo
Sandstone landscape. The Rugged Veld on Basalt is the most dominant landscape and occurs in
various forms on the reserve. Colophospermum mopane and Combretum apiculatum form the
dominant woody vegetation which falls within the Mopane-Combretum shrub-savanna plant
community (Joubert 1984; Nchunga 1978). The Floodplain landscape occurs along the Limpopo,
Shashe and Motloutse rivers and is made up of mostly alluvial soils with Acacia tortillis savannah
(McKenzie 1990). Acacia albida Gallery forest and Croton megalobotrys thickets also occur in
the riverine areas fringed by Salvadora angustifolia/Acacia tortilis Bushveld (Joubert 1984).
Boscia foetida savannah is found on the old undulating floodplains (McKenzie 1990). The
Sandstone landscape is found only in the south and west of the reserve, it is composed of
sandstone outcrops and sandveld valley composed of sparse woody vegetation and grasses
(McKenzie 1990). The vleis are dominated by stands of tall Elephant grass, Sporobolus
consimilis, which can reach a height of 1.5 to 2m, woody vegetation is absent in this habitat
(Joubert 1984). Figure 2.6 shows the main vegetation types found on NOTUGRE together with
the water channels and waterholes.
2.8 Fauna
NOTUGRE supports large populations of ungulates particularly in the dry season due to the
presence of both natural and artificial water points. Common large and medium mammal species
are listed in Appendix I. The reserve has resident populations of all naturally occurring large
carnivores, with the exception of the African wild dog which only occurs sporadically. Wildlife
is free to move out of the reserve into neighbouring properties where permitting.
The following species used to occur in NOTUGRE but have been absent for over 40 years (Lind
1974):
26
Alcelaphus buselaphus Red hartebeest
Damaliscus lunatus Tsessebe
Hippotragus equinus Roan antelope
Hippotragus niger Sable antelope
Oryx gazella Gemsbok
Syncerus caffer African Buffalo
Redunca arundium Common reedbuck
A possible reason for the local extinction of these species is the change in vegetation particularly
along the river banks which are believed to have been more densely vegetated (Lind 1974).
Changes to the environment such as artificial waterholes and human activity brought about by
settlers may also have influenced the movement, concentration and general wildlife populations
(Lind 1974).
2.9 Study animal
Accounts of cheetahs in NOTUGRE are scarce. Between 1966 and 1971 only two sightings of
two cheetahs were recorded (Walker 1971) and there were reports of a number (number
unspecified) of cheetahs poached and found at a trading store south of the Motloutse River
(Walker 1971). Lind (1974) describes the cheetah as rare and only seen sporadically in
NOTUGRE. He gives a few accounts of cheetah sightings and broadly estimates the population
to have a total of seven individuals of unspecified ages in 1972 and nine individuals by the end
of 1973. Numbers were estimated based on total number of sightings and not individual
recognition. Lind (1974) described the species as at risk of being locally extinct and
recommended strict protection by keeping disturbance to a minimum. Prior to my study, there
had not been any formal assessment of the NOTUGRE cheetah population.
27
CHAPTER 3 TRAPPING THE ELUSIVE CAT: USING INTENSIVE CAMERA TRAPPING
TECHNIQUES TO ESTIMATE THE DENSITY OF CHEETAHS IN THE NORTHERN
TULI GAME RESERVE, BOTSWANA
3.1 Introduction
Large carnivores play critical roles in the functioning of ecological systems and often act as
umbrella species for the maintenance of biological diversity (Ferreira & Hofmeyr 2014; Ripple
et al. 2014). Additionally, cheetahs (Acinonyx jubatus) and other large carnivores can be regarded
as flagship species, providing revenue for eco-tourism operations (Buk & Marnewick 2010;
Ferreira & Hofmeyr 2014; Ripple et al. 2014). However, many large African carnivores have
disappeared from their historical ranges due to habitat fragmentation, prey depletion and direct
persecution, and their persistence relies mostly on the success of conservation strategies (Ray et
al. 2005). To implement conservation actions effectively it is essential to have a reliable
understanding of the status of resident carnivore populations (Carbone et al. 2001).
Little is known about the status of cheetahs in Botswana and there is no estimate of the size of
the cheetah population of the Northern Tuli Game Reserve (NOTUGRE). McKenzie (1990)
vaguely refers to cheetah numbers being very low prior to 1984 but with a notable increase
thereafter. Given the critical conservation status of cheetahs (IUCN 2013), it is important to have
reliable population estimates to adequately evaluate the success of conservation efforts (Durant
et al. 2007). Furthermore, the unique location of NOTUGRE, adjacent to two international
borders (with South Africa and Zimbabwe), emphasizes the importance of having accurate
population estimates for sound managerial decisions to be made throughout the cheetahs’ range,
irrespective of geopolitical boundaries.
Many aspects of cheetah ecology make it extremely difficult to monitor their populations. They
occur at very low densities, and are elusive, cryptic, and highly mobile (Gros 1998; Marnewick
et al. 2006, 2008; Durant et al. 2010). Cheetahs sometimes aggregate at smallscale, local transient
hotspots (for example in areas with high prey densities and low predator densities), that may be
miss-extrapolated to large-scale high cheetah density (Durant 1998; Durant et al. 2010).
Additionally, the large home ranges of cheetahs may give the false impression of high cheetah
numbers due to repeat sightings of the same individual(s) at several locations over large areas
28
(Marker et al. 2008a; Houser et al. 2009; Buk & Marnewick 2010). Direct counts of cheetahs are
logistically impractical and incur high financial and time costs making them rarely feasible
(Durant et al. 2007; Balme et al. 2009). Population sizes can, however, be estimated using
indirect methods (Thompson et al. 1998; Karanth & Nichols 2002; Balme et al. 2009). For
example, the survey of animal signs, public interviews and photograph submissions, or inferences
of population densities from indices such as prey biomass and habitat suitability (Karanth et al.
2010).
The method selected to estimate numbers should consider the species, the area and habitat, the
available budget and the amount of skilled manpower and time available. However, the objectives
of the study should be of utmost importance (Karanth & Nichols 2002; Henschel & Ray 2003).
The objectives of a study may vary from a simple presence-absence survey, to relative abundance,
to absolute abundance and density estimates (Henschel & Ray 2003). In addition, the detection
probability, which is the probability of an animal being included in the count statistic, should be
considered regardless of the method such that the sample size or number of target animals
detected is sufficient for sound population estimates (Nichols 1992; Karanth & Nichols 2002).
3.1.1 Camera trapping
Remotely-triggered camera trapping is a non-invasive method for monitoring rare, cryptic
mammals (Carbone et al. 2001). It can be successfully used to systematically survey individually
identifiable big cats (Soisalo & Cavalcanti 2006). Individuals can be identified by unique natural
markings such as spot or stripe patterns, which allows for population estimates to be calculated
by capture-recapture methods (Otis et al. 1978). Photographs from the surveys provide encounter
history data, representing the sequence of individual observations generated from camera traps,
with occasions and spatial locations of individual photo captures.
This method has been successfully used to provide population estimates for a number of
individually recognizable felid species (see Table 3.1). Although camera-trapping studies of
cheetahs have been completed in north-central Namibia (Marker et al. 2008b) and the
Thabazimbi district of the Limpopo Province of South Africa (Marnewick et al. 2008), both
studies used fewer than 13 sampling locations. O’Brien & Kinnaird (2011) also published
abundance estimates of cheetahs using camera traps but the four positive returns were insufficient
to derive a reliable population estimate. These are apparently the only published cheetah
population estimates using camera-trapping and therefore this study represents the first intensive
camera-trapping survey in Africa to estimate absolute abundance and density of cheetahs.
29
Table 3.1 Examples of camera trapping studies for individually recognisable felid species
Species Studies Tigers Panthera tigris Karanth 1995; Karanth & Nichols 1998, 2002;
Carbone et al. 2001 Jaguars Panthera onca Silver 2004; Soisalo & Cavalcanti 2006; Kelly et al.
2008; Negrões et al. 2012; Noss et al. 2013 Ocelots Leopardus pardalis Trolle & Kéry 2003; Maffei et al. 2005, Dillon &
Kelly 2007 Snow leopards Panthera uncia Jackson et al. 2010 Leopards Panthera pardus Henschel & Ray 2003, Balme et al. 2009; Gray &
Prum, 2012; Borah et al. 2013 Pumas Puma concolor Kelly et al. 2008; Negrões et al. 2010
3.1.2 Capture-recapture method
Closed-population capture-recapture models are typically applied to estimate the relative
numbers (or density) of cryptic carnivores (Nichols 1992). For the application of this method,
two assumptions need to be met; 1. The population is demographically closed; and 2. Individuals
cannot have zero probability of capture (White et al. 1982; Nichols 1992; Karanth & Nichols
2002). Population closure is practically met by using a survey period that is sufficiently short that
it is unlikely that deaths, births, immigration or emigration will occur during the surveyed period
(Otis et al. 1978; Karanth & Nichols 2002; Tobler & Powell 2013). However, the capture
probabilities for cheetahs, especially in semi-arid habitats with lower prey density, are expected
to be low (Gros et al. 1996; Marker et al. 2008b; Buk & Marnewick 2010).
Therefore, a balance needs to be found where survey length is short enough to satisfy population
closure, but long enough for sufficient data to be collected for population estimation (Tobler &
Powell 2013).
It is also essential that individuals of the target species be reliably distinguished from each other
throughout the study (White et al. 1982). Cheetahs are individually recognisable by their unique
spot patterns (Kelly et al. 1998) and image quality and trap placement are therefore factors which
must be carefully considered (Karanth & Nichols 2002). Population density provides a useful and
comparable population statistic (O’Brien & Kinnaird 2011). The density of a population is
defined as the number of adult animals averaged across the study area and is typically expressed
as the number of animals per 100 square kilometers (Karanth & Nichols 1998). To calculate
density the effective area sampled needs to be known and is estimated by adding a buffer around
the trap array (Karanth & Nichols 2002; Soisalo & Cavalcanti 2006). The effective area sampled
30
is conventionally calculated using ad hoc approaches whereby the mean maximum distance
moved (MMDM) or half of the mean maximum distance moved (HMMDM) by the animals being
studied is calculated (Otis et al. 1978; White et al. 1982; Karanth & Nichols 1998, 2002; Soisalo
& Cavalcanti 2006). This approach has been heavily criticized because the effective trapping area
(ETA) varies considerably with the chosen buffer strip method, and consequently influences
density estimates (Efford 2004; Soisalo & Cavalcanti 2006; Borchers & Efford 2008; Foster &
Harmsen 2012; Gerber et al. 2012; Gopalaswamy et al. 2012; Noss et al. 2013; Tobler & Powell
2013).
A relatively novel approach has been developed using Spatial Explicit Capture-Recapture
(SECR) models (Efford 2004; Borchers & Efford 2008; Royle et al. 2009a). SECR models
incorporate the geographic locations of camera traps and the individual animal captures within
the trap array, thereby accounting for unequal detection probabilities among individuals and
enabling direct estimates of population size and density (Borchers & Efford 2008; Royle et al.
2009b; Sun et al. 2014). The SECR method calculates individual specific detection probabilities
by estimating the activity centres of individuals and the camera trap locations (Borchers 2010;
Sun et al. 2014). Furthermore, SECR models allow for non-regular trap locations while still
providing precise estimates of abundance (Sun et al. 2014). SECR methods have consequently
become the preferred method for calculating population estimates from camera-trapping data and
have been implemented for several recent camera trap surveys of individually identifiable large
carnivores (Royle et al. 2009a; Gardner et al. 2010; Kalle et al. 2011; Sollmann et al. 2011; Grant
2012; Foster & Harmsen 2012; Gerber et al. 2012; Mondal et al. 2012; Noss et al. 2012, 2013;
Gray & Prum 2012; Tobler et al. 2013).
3.1.3 Objectives
In this chapter, the cheetah population density of NOTUGRE is estimated using camera trapping
techniques and SECR analyses. The influence of placement of camera traps at scentmarking trees
on cheetah capture rates is also investigated. Finally, the effect of survey duration on sample size
and resulting population estimates for a carnivore species that occurs at a low population density
is explored.
3.2 Methods
Two camera-trapping surveys were carried out using two different trapping arrays. The first
survey followed the more traditional approach of having camera traps set uniformly over the
landscape in a systematic pattern (Otis et al. 1978). The second survey had camera traps placed 31
at sites presumed to increase the probability of capturing cheetahs, resulting in an irregular pattern
of trap locations across the study area (Marker et al. 2008b). Both surveys were conducted in the
centre of NOTUGRE on five different properties, including Mashatu, Fika Futi, Naledi, Kanda
and Uitspan North, and covered approximately 240 km² (Figure 3.1). The location was chosen
for practical purposes but also to avoid the edges of the reserve where theft of cameras may have
been a problem.
Figure 3.1 A map of NOTUGRE illustrating the properties included in the study area (green polygons) for the camera trapping surveys.
32
3.2.1 First Survey – Regular trap configuration
Twenty Cuddeback Attack (Non Typical, Inc., Green Bay, WI, USA) camera traps were used at
60 locations within NOTUGRE (Figure 3.2). A stratified, random sampling technique was used
(Otis et al. 1978; White et al. 1982; Thompson et al. 1998; Borah et al. 2013), deploying camera
traps in the best locations, typically along trails or other well-travelled animal paths (i.e. the
locations most likely to capture moving animals) within a buffer. This ensured an even sampling
effort across the landscape and an equal detection probability for all individuals, reducing
sampling biases from spatial variation in capture probabilities (Karanth & Nichols 1998; Foster
& Harmsen 2012).
Based on cheetah movements observed in the study area, a grid with equally spaced points at 3.7
km intervals was placed over a map of the surveyed area using ArcMap 10 (ESRI, Redlands, CA,
USA). These predetermined points represented ideal camera trap placements, but actual camera
traps were set within 200 m (mean distance and standard deviation = 164 ± 94 m) of the
predetermined points, thus had a tolerance of 4.4 ± 2.5%. Camera traps were placed within this
buffer zone at sites presumed to maximise the likelihood of photographing a moving animal,
usually on well-defined animal paths (Balme et al. 2009). The 3.7 km spacing between the units
was chosen to ensure that no cheetah would go undetected (Karanth & Nichols 1998, 2002).
Studies designed to estimate the abundance of a species require that camera traps be placed such
that the entire area sampled does not have any large gaps in which a cheetah’s movements could
go undetected during the sampling period (Karanth & Nichols 2002). In other words, no cheetah
has a capture probability of zero. The spacing between camera traps is typically based on the
average home range or minimum home range of the target species (Karanth & Nichols 2002).
However, the size of cheetahs’ home ranges varies substantially among geographical locations
and social groups (Gros et al. 1996; Broomhall et al. 2003; Bissett & Bernard 2007) and there
were no prior data on cheetah home ranges for the study area or surrounding areas. Information
about the movement of a resident adult female cheetah with sub-adult cubs was available and
was used to calculate the average daily distance moved. The movement data were obtained from
global positioning co-ordinates taken every four hours from a satellite collar fitted to the cheetah
(E. Brassine, unpublished data). The cheetah was collared by a qualified veterinarian for routine
monitoring. The daily distance travelled was calculated by adding the distance between
consecutive locations within a day (Hunter 1998). This average daily distance moved was used
to calculate the minimum distance between camera trap sites.
33
Figure 3.2 The locations of camera traps (n = 60) for the first survey using a systematic grid method. (ArcMap 10; projected: Transverse-Mercator, spheroid W GS84, central meridian 29; map units: meters).
Using two opposed cameras per station is preferable to capture both sides of the animal for
individual identification and to increase the detection rate (Karanth & Nichols 2002; Negrões et
al. 2012). However, only one camera was used per station in my survey so that more traps could
be deployed over a larger area, thereby increasing the number of independent locations and
maximising the chances of detecting every cheetah in the area (Foster & Harmsen 2012).
When surveying rare or sparse species it is best to sample broadly across the study area as this
increases the likelihood of captures (Foster & Harmsen 2012). Typically, camera-trapping
surveys are conducted over a short period to ensure demographic closure (Karanth & Nichols
2002; Royle et al. 2009a). My survey was carried out over a 90 day period during the hot/wet
season (December – March 2013). Due to the large size of the survey area (±240 km²) and the
limited number of cameras (n = 20), the Adjacent Block method (Karanth & Nichols 1998, 2002)
was implemented to ensure that the whole sampling area was covered. The sampled area was
divided into three sections and each section was sampled sequentially for approximately 30
continuous days (Karanth & Nichols 2002). Cameras were collected from their first location in
one section and deployed to their new location as quickly as possible (approximately three days
to move all cameras). The total number of days that cameras were active is the duration of the
survey, with each day (24-h period) defined as a sampling occasion, starting at 12h00 and ending
at 11h59 (Otis et al. 1978), when cheetahs are believed to be least active (Hayward & Slotow
2009).
34
The cameras were set to take high quality (5MP) images and the strobe flash range was set at 30
feet (9.14 m). This was occasionally reduced to 10 (3.04 m) or 20 feet (6.09 m) when an animal
was likely to come closer to the camera so as to reduce the risk of overexposed images. The
cameras used four D-cell batteries, a 4GB SD card and a passive infrared sensor to detect heat
and motion. The cameras were housed in steel protective casings and fastened to trees. Chains
and padlocks were also used to secure the cameras against theft. Cameras were secured
approximately 0.3 m above the ground and were active 24h/day with a 1 minute delay between
consecutive photographs to minimize unnecessary captures of gregarious, non- target species.
The cameras were inspected, on average, every 15 days to replace batteries and memory cards
and to ensure that they were operating normally. No baits were used at camera trap stations to
prevent heterogeneous capture probabilities (Foster & Harmsen 2012). However, no effort was
made to conceal human scent.
All data from camera traps were summarized in a comprehensive spreadsheet. The number of
active days, or trap-days, was calculated for each station. Every day that a camera was active was
deemed one active day. If cameras malfunctioned, had technical problems (such as no flash
triggered at night or flat batteries), or were damaged by elephants (Loxodonta africana) or
flooding, those days were excluded from the data analyses. Thus, active days included only
problem-free days. Independent photographic events were defined as consecutive photographs
of the same species taken more than one hour apart, or non-consecutive photographs of
individuals of the same species (Tobler et al. 2008). The number of independent events per 100
trap days [relative abundance index (RAI)] was calculated for each species (Karanth & Nichols
2002).
3.2.2 Second survey – non-random configuration using scent-marking posts
The probability of detection is a fundamental aspect that needs to be carefully considered in order
to obtain robust estimates of population size (Long et al. 2008). A large enough sample size relies
on the capture probability of the species being studied (Otis et al. 1978), which depends on a
number of variables, such as survey design, habitat type, prey availability, and most crucially on
the behaviour of the target species (Soisalo & Cavalcanti 2006). In this study an important
behavioural trait was cheetahs’ communication with conspecifics through scentmarking (Eaton
1970; Marker et al. 2010; Soso et al. 2014). Scent-marking can take the form of defecation on or
under a tree, urine spraying, and clawing (Eaton 1970; Marnewick et al. 2006; Soso et al. 2014).
Trees are predominantly used for scent-marking posts but rocks, termite mounds and even man-
made objects may also be used (Eaton 1970). Careful choice of trap location may increase the
35
probabilities of capturing the target species, and hence produce a more accurate representation of
the true population at the study site (Soisalo & Cavalcanti 2006).
The second camera trapping survey used known scent-marking posts for camera trap locations
(Marnewick et al. 2006, 2008; Marker et al. 2008b). Field guides working in NOTUGRE have
observed cheetahs using scent-marking posts and, with their assistance, a total of 104 such sites
were identified and mapped as potential trapping locations. A proximity test was run in ArcMap
10 to calculate distances between all scent-marking posts and data were cleaned; effectively
removing scent-marking posts that were within 250 m of other scent-marking posts. Where more
than one scent-marking post lay within a selected area, the site with the most recent signs of
cheetah activity (presence of scats, urine spray, and tracks) and with the least human interference
would be selected. Accordingly, 60 camera trap placement sites were chosen (Figure 3.3); with
the number of sampling points consistent with the first survey.
The furthest spacing between scent-marking posts (3.13 km) fell within the chosen required
maximum distance between camera trap placements (3.7km). This ensured that there were
probably no gaps sufficiently large to contain a cheetah’s movements within the sampled area
and that all cheetahs had a non-zero detection probability (Karanth & Nichols 2002).
Figure 3.3 Camera trap locations (n = 60) at identified scent-marking posts for the second camera trap survey. (ArcMap 10; projected: Transverse-Mercator, spheroid WGS84, central meridian 29; map units: meters).
Cameras were set in the cheetah’s anticipated path to photograph the flank of the animal because
broadside images facilitate easier identification (Marnewick et al. 2006). Where possible, brush
36
was packed around the scent-marking tree leaving only one access point to encourage the animals
to move in front of the camera (Marnewick et al. 2006). A combination of Cuddeback Attack (n
= 24) and Bushnell Trophy CamTM IR (Bushnell Outdoor Products, Overland Park, Kansas,
USA) (n = 6) camera traps were used. Cameras were only operational at 30 locations during any
given sample occasion. Thus, the Adjacent Block method was implemented, with the sampled
area divided into two blocks and a camera rotation after 45 consecutive days to cover the entire
sampled area. The Bushnell cameras were set to take a burst of three photographs per trigger to
aid in identification, but for every trigger event, consecutive photographs were recorded as a
single capture. Cuddeback Attack cameras allowed for a short video clip (30 seconds) to be taken
after each daytime trigger event. This function was activated to aid individual cheetah
identification. All other camera settings and positioning were as per the first camera trapping
survey.
The survey ran for 90 days and was carried out during the cool/dry season (June –September
2013). Cameras were checked approximately every two weeks with an initial check after three
days to ensure that the camera was operating correctly and was properly positioned to maximise
the chances of captures.
3.2.3 Extended survey
Small sample sizes are typical of capture-recapture studies for carnivores that have large home
ranges (O’Brien & Kinnaird 2011). Nonetheless, the dataset needs to include captures and
recaptures of a sufficient proportion of the population to calculate effective sampled area and
density (Foster & Harmsen 2012). A larger sample size, and hence precision, can be obtained by
adapting the design of the survey, this includes using species-specific targeted placement;
increasing the number of sampling points; using a larger sampling area; increasing the density of
trapping points within the sampled area; and extending the duration of the sampling period (Otis
et al. 1978; O’Brien & Kinnaird 2011).
To increase the number of captures, the survey period of the second camera trapping survey was
extended after the initial 90 day survey. The 30 camera traps were left at their position for a
further 40 days, extending the number of trapping days to a total of 130 days. While a long survey
period may be necessary for species with low detectability to have sufficient captures for analyses
(Foster & Harmsen 2012), the assumption of demographic closure may become violated (Foster
& Harmsen 2012). Thus, population closure tests and SECR analyses were performed using all
cheetah photographic captures over the extended sampling period. Density estimates for the two
different sampling period lengths were compared using a Student’s t-test. 37
3.2.4 Cheetah identification
Cheetah photographs were categorized and analysed with Adobe Photoshop Lightroom 3.6. All
photographic captures of cheetahs were analysed by visual inspection of spot patterns to
determine the identity of each cheetah and each individual was given a unique identity number
(Kelly et al. 1998; Kelly 2001). The identification of individuals and capture events was based
on the guidelines below (Caro 1994; Karanth 1995; Heilbrun et al. 2003).
Individuals were identified based on spot patterns or individual spots on the body, tail, legs and
face. At least two, but preferably three, unique features or human-made markers (e.g. a collar)
were required to identify an individual. One different feature was considered sufficient to
consider that two photographs represented two different individuals. Photographs of poor quality,
or where spot patterns were obscured, were marked as unidentifiable and excluded from the
analysis. A photograph was considered to be a first capture if it could not be matched with any
individuals in previous, older photographs. Re-captures were photographs depicting an individual
already identified. All individuals were sexed based on presence/absence of scrotal testes.
The photographs were independently analysed by two people to ensure their correct classification
(Kelly et al. 2008). If an individual’s identity could not be agreed upon, these photographs were
excluded from the analysis. A cheetah identikit, developed during the photographic survey (see
Chapter 4 and Appendix III), was used to assist with identification. Only adult cheetahs were
considered for analysis of population estimates. The sampling occasion, time, location, and
individual cheetah identity of each capture event were recorded in a spreadsheet. Capture
histories were prepared for each adult identified in the camera trapping survey with a sampling
occasion defined as 1 day (24 hours) starting at 12h00.
3.2.5 Data analysis
Tests for population closure were performed using the CloseTest program version 3. The program
tests capture-recapture data for closure using two tests (Otis et al. 1978; Stanley & Burnham
1999). There are a number of programs that are available for calculating population estimates
using capture-recapture data such as SPACECAP and DENSITY (Efford et al. 2004;
Gopalaswamy et al. 2012). These are the most commonly used programs for running SECR
models, with each program using a different (and therefore independent) approach for running
the analysis. SPACECAP uses a Bayesian modelling framework and DENSITY uses maximum
likelihood-based approach (Gopalaswamy et al. 2012). Although analyses by the program
SPACECAP take much longer to run than DENSITY, it was preferred as it allowed for inference 38
about the locations of individuals that were not photographed during the survey and could thus
be used for modelling demographically open populations (Gopalaswamy et al. 2012). Another
advantage of SPACECAP is that the Bayesian framework offers non-asymptotic inferences
which are applicable for small data samples typical of camera trapping studies of carnivores that
occur at low densities (Gopalaswamy et al. 2013).
Density estimates are calculated in SPACECAP using information on capture histories in
combination with the distribution of individuals (trap sites) and each traps’ active days (dates
when camera trap locations were active and operational), providing more accurate, precise, and
hence more reliable results (Gopalaswamy et al. 2012). The model firstly determines an
individual’s activity centre and then estimates the density of these activity centres across a
precisely defined area containing the trap array (Gopalaswamy et al. 2013). Furthermore, the
models consider the traps as functioning independently and this allows individuals to be captured
in multiple traps during a capture occasion and even multiple times by the same camera trap,
which is realistic in camera-trapping studies (Royle et al. 2009a).
SPACECAP runs as a package in the program R version 3.0.2 (R Development Core Team)
(Gopalaswamy et al. 2013). SECR analysis in SPACECAP requires specific input files.
Three input files are required; these files consist of the following:
• Animal capture detail
• Trap deployment detail
• State-space detail
Guidelines for creating the three input files can be found in the SPACECAP manual
(Gopalaswamy et al. 2013). Spreadsheets were created using Microsoft Excel and the input files
were saved in an ASCII comma separated values (.csv) format in the working directory. All X
and Y-co-ordinates must be expressed in the Universal Transverse Mercator UTM projection
system for computation in SPACECAP.
The third file (state-space detail) requires the creation of potential activity centres within the
state-space. The state-space or ‘S’ represents the surveyed area containing the camera traps
combined with an extended area surrounding it. The state-space is represented by a fine grid of
equally-spaced points that represent all possible activity centres (or home range centres) of all of
the individuals in the population surveyed (Gopalaswamy et al. 2013). Point spacing of 500 m is 39
commonly used in the point array but because of the relatively large home ranges of cheetahs
1000 m point spacing was selected. The distance between points should be such that 10-20 points
might lie in a single home range of an individual (A. Royle, research statistician and author of
SPACECAP, pers. comm.). Potential home range centres were generated using ArcMap10 in
conjunction with the Repeating Shapes for ArcGIS extension Tool (Jenness 2012). The state-
space requires being sufficiently large to ensure stability in the density estimate, which usually
requires a buffer strip to be added to the trap array that is two or three times larger than the
encounter probability parameter (Gopalaswamy et al. 2013). A “Minimum Area Rectangle” is
formed by connecting the outermost camera trap locations in a rectangle and a buffer is created
around this minimum area rectangle.
3.2.6 Buffer
The buffer region should be sufficiently large for individual animals outside the buffered region
to have zero probability of being photo-captured by camera traps during the survey. For the
analysis of the data, the state-space boundaries were calculated using three different buffered
distance methods (see Figure 3.4): Double the diameter of the minimum known home range size
for cheetahs (11 km²; Purchase & du Toit 2000) (Buffer width = 3.74 km). The diameter of a
known home range for that specific site (E. Brassine, unpublished data), approximating the home
range as a circle (Buffer width = 8.97 km). The Maximum Distance Moved (MDM) (Buffer width
= 28 km) – the centre point of the home range of a cheetah fitted with a satellite collar was
calculated by averaging all of the GPS co-ordinates. The furthest fix from this centre point was
used to measure the MDM. If home range data from more than one collared cheetah had been
available, the average maximum distance moved would have been used to calculate the buffer
distance, as sample size could affect this measurement.
The adequacy of each model is evaluated based on its Bayesian posterior probability (Pvalue). A
model that provides an adequate description of the data will have a Bayesian Pvalue near 0.50,
extreme values (near 1 or 0) indicate that the model is inadequate (Gopalaswamy et al. 2013).
The habitat suitability indicator column required in the third input file was created with data from
Google Earth. Aerial imagery of the state-space area was used to indicate areas unsuitable for
cheetahs. Selected by the author, unsuitable areas included human settlements, large water
bodies, fenced agricultural farms (farms with high human activity and maintained game fences),
and mining areas (Pettorelli et al. 2009; Gopalaswamy et al. 2013). Home range centres that fell
on these areas were identified as locations where cheetahs could not exist and marked with a ‘0’
next to their co-ordinates. Regions of suitable habitat were described by a grid of equally spaced 40
points representing 1 km² over the state-space. The activity centres are assumed to be uniformly
distributed over this area of suitable habitat.
The SPACECAP input files were uploaded and appropriate model combinations were chosen for
analysis (Gopalaswamy et al. 2013). The following model definitions were selected: trap
response absent, spatial capture-recapture, and detection function was set to half-normal
(Gopalaswamy et al. 2013). The Markov-Chain Monte Carlo (MCMC) parameters were set to
the recommended default values: 50 000 iterations, 1000-sample burn-in, no thinning was
selected (value of 1) and data augmentation of 35 was chosen. To analyse the complete data
model (the model with a fixed number of activity centres) where the number of animals in the
population is unknown, the method of data augmentation is used. The data augmented must be
sufficiently many that the posterior probability distribution of N is not truncated. Following the
recommendation by Royle et al. (2009b) that the data to be augmented should be five to ten times
the number of identified individuals, data augmentation was set to 35 (five times seven). The data
augmentation value represents the maximum allowable number of possible animals within the
state-space (Gopalaswamy et al. 2012). The behavioural response was not chosen as baits or lures
were not used in the survey, thus an individual’s encounter probability before and after the initial
encounter was expected to be similar. Movement of individuals was non-random in this case as
individuals will use certain scent marking posts within their home range (Caro 1994).
Figure 3.4 An example of the spatial data created in ArcMap 10 for the third input file “Potential Home Range Centres” for the program SPACECAP showing the state-space boundaries for three different buffered distances.
41
3.3 Results
3.3.1 First Survey – Regular trap configuration
A total of 1616 active days were logged during which 3346 animal photographs were taken and
only nine (0.27%) were photographs of cheetahs. Cheetah photographs were recorded at only two
of the 60 sampling locations and all but one of these events occurred at a camera trap station that
had been placed at a known cheetah scent-marking post. From the photographs 32 mammal
species and 23 bird species were identified and no reptile species were captured. Relative
abundance indices (RAI) and the proportion of total photographs taken are shown in Appendix
II for all recorded mammal species. The most common mammal species, based on capture
frequencies (CF > 2.0), were impala (Aepyceros melampus), followed by elephant, giraffe
(Giraffa camelopardalis) and eland (Tragelaphus oryx). Eleven predator species were identified
and the most frequently photographed were spotted hyena (Crocuta crocuta), blackbacked jackal
(Canis mesomelas), and leopard. The least common species were bushpig (Potomachoerus
porcus), lion (Panthera leo), banded mongoose (Mungos mungo), and bushbuck (Tragelaphus
scriptus) (all photographed only once). No further analyses to assess cheetah population size were
carried out due to the insufficient number of cheetah captures.
3.3.2 Second survey – non-random configuration using scent-marking posts
The second study had a total of 2660 active camera trapping days and of the 3323 animal
photographs, 53 (1.6%) were of cheetahs captured at 11 of the 60 camera trap sampling locations.
Forty-nine species, including 28 mammal and 21 bird species, were recorded. Appendix III shows
the RAI and proportion of total photographs taken for the entire mammal species recorded. This
survey detected two mammal and nine bird species which were not recorded in the first survey.
However, five mammal and 11 bird species which were captured in the first survey were not
recorded by the second survey. Anthropogenic activity was high (4.4%; n = 116 photographs),
because many of the scent trees were placed on hills used as stopping points during game viewing
drives.
Cheetah photographs made up 18 independent capture events. A capture event includes all
photographs of an individual within a 24 hour activity period at a camera station (O’Brien et al.
2003; Gerber et al. 2012). A total of seven adult cheetahs were identified from photographs (two
females; five males) and five cheetah photographs were excluded from the analysis as the
individuals could not be identified. However, the cheetahs in these photos were cubs and would
42
have been excluded from the analysis regardless of whether identification was possible or not.
Details of individual cheetah visits, capture location and capture occasion are shown in Table 3.2.
Table 3.2 All capture and re-capture details of individual cheetah visits (sample occasion)
recorded from the second camera trapping survey in the Northern Tuli Game Reserve, Botswana.
Independent capture events used for analyses are not shown.
Sample
Occasion*
Time Location ID Number of photographs
Cheetah ID
9 11:46 24 1 CM6 10 12:07 25 1 CM5; CM6 38 10:49 25 1 CM3 45 11:14 18 1 CM2 46 14:31 22 2 CM1; CM2 46 15:48 22 4 CM1; CM2; unidentifiable 50 50 13:20 58 3 CF3; 2 cubs 63 06:33 54 3 CF3; 3 cubs 72 06:33 56 5 Cub 78 05:18 33 3 CM5; CM6; unidentifiable 85 87 10:43 40 6 CF4 87 19:34 58 2 CM2; CM3 87 04:46 58 3 CM1 87 04:51 58 2 CM2 87 07:28 57 1 CM1; CM2; CM3 88 07:45 56 2 CM2; CM3 88 20:20 55 9 CM1; CM2; CM3 88 05:09 55 1 CM1 88 05:26 54 3 CM1; CM2; CM3
*Sample occasion refers to the day on which cheetahs were captured within the survey period with sampling occasion 1 referring to the first day of sampling.
Capture frequency ranged from one to five per individual, with an average of 2.86 captures per
individual. The number of photographs per sampling occasion ranged from one to nine, with an
average of 2.79 per sampling occasion. Latency or time delay to first photograph for each
individual ranged from 9 to 85 days.
3.3.3 Extended survey
A further 1090 days were logged from the extended survey period which resulted in a total of
3750 recorded active camera trapping days (Table 3.3). An additional seven camera trap locations
photo-captured cheetahs, which accounted for a total of 13 events including 147 cheetah
photographs, increasing the total sample size from 18 to 31 capture events at a total of 18 camera
trapping locations (Table 3.3). No new individual cheetahs were recorded during this extended
survey. However, capture frequency ranged from two to 10 per individual, with an average of
5.57 captures per individual. Capture details of individual cheetah visits are shown in Table 3.4. 43
Table 3.3 Summary of the first, second and extended camera trapping surveys conducted in
NOTUGRE.
First survey Second survey Extended survey No. of active camera-trapping days 1616 2660 3750 Total number of photo-captures 3346 3323 4823 Cheetah photo-captures 9 53 200 Cheetah capture events 5 18 31 Number of individual cheetah identified 2 7 7 Number of sampling locations that captured cheetahs
2 11 18
Capture frequency per individual (mean) N/A 2.86 5.57
Table 3.4 Capture details of individual cheetah visits (sample occasion) recorded during the
extended second camera trapping survey in the Northern Tuli Game Reserve, Botswana.
Sample occasion
Time Location ID Number of photographs
Cheetah ID
101 1:58 33 2 CM6 101 9:13 44 32 CM5; CM6 124 11:41 51 1 CF4 126 6:20 47 3 CF8 (cub) 130 2:37 52 3 CM1; CM2; CM3 130 3:16 51 2 CM2; unidentifiable 130 3:36 33 2 CM6; unidentifiable 130 3:49 47 3 CM2; unidentifiable 130 7:43 41 9 CM2; CM3 131 23:03 43 1 CM5 131 3:42 38 3 CM1; CM2; CM3 133 19:18 33 1 CM5; CM6
3.3.4 SECR analysis for the second camera trapping survey
Closure tests were inconclusive due to the small dataset that only had a few individual captures
and recaptures. Small sample sizes and unequal capture probabilities can negatively affect closure
tests (Otis et al. 1978; Soisalo & Cavalcanti 2006). However, previous studies of large felids
have indicated that a three-month sampling period is sufficient to meet the closure assumption
(Karanth 1995; Karanth & Nichols 1998). Density estimates were sensitive to the buffer width
estimator, with density estimates increasing with a decreasing state-space area (Table 3.5).
Cheetah density over the state-space with a buffer of 28km was estimated at 0.55 cheetahs/100
km² with a 95% confidence interval of 0.24 (lower level) and 0.90 (upper level). However, a
buffer width of 8.97 km showed considerable difference with an estimated density of 1.24
cheetahs/100km² and the smallest buffered distance used (3.74km) for the state space area
estimated cheetah density at its highest with estimates of 1.69 cheetahs/100km². Bayesian P-
values for the different models suggest that all models may be appropriate (Table 3.5).
44
Table 3.5 Summaries of the Bayesian SECR analyses using three different buffer width
estimators to create the state space area. The Bayesian P-value gives the adequacy of each model.
Buffer Width (km) Variables* Mean SD 95% Lower HPD level
95% Upper HPD Level
Bayesian P- value
28 km buffer; state space area of 4708 km²
sigma 6.17 1.63 3.57 9.60 0.46 lam0 0.02 0.01 0.00 0.04
Psi 0.60 0.22 0.24 1.00 Nsuper 25.39 9.09 11.00 42.00 Density 0.54 0.19 0.24 0.90
8.97 km buffer; state space of 1166 km²
sigma 4.98 1.37 2.78 7.59 0.56 lam0 0.01 0.00 0.00 0.02
Psi 0.35 0.16 0.10 0.67 Nsuper 14.22 6.27 7.00 27.00 Density 1.24 0.55 0.61 2.36
3.74 km buffer; state space of 591 km²
sigma 4.79 1.12 2.90 6.96 0.58 lam0 0.00 0.00 0.00 0.01
Psi 0.25 0.09 0.09 0.43 Nsuper 9.90 2.72 7.00 15.00 Density 1.69 0.46 1.19 2.56
* Sigma (σ) represents the range parameter of an animal; lam0 (λ0) is the expected encounter rate and can be used to estimate capture probability; parameter psi (ψ) represents the proportion of the actual number of animals and the maximum allowable number which was set during data augmentation; Nsuper is the population size of individuals for the prescribed state-space; density is calculated from the estimated number of activity centres located in the state-space and is expressed as individuals per 100 km².
3.3.5 Model refinement
Following recommendations by A. Royle (pers. comm.), the author of SPACECAP, the MCMC
parameters were changed to the following: number of iterations 100 000 and burn-in values of
2000 generations to accommodate for the small sample size and large movements observed in
the dataset. Buffer widths of 20, 25 and 30 km were tested and posterior summaries were used to
estimate when the density estimates would stabilize. Based on this approach, the 28 km buffer,
forming a 4708 km² state-space was considered to be the most appropriate method for calculating
the buffer width. Another SECR analysis was run in SPACECAP using this buffer width with
200 000 iterations and 4000 burn-in generations as MCMC parameters, higher MCMC
parameters were chosen to ensure that the key parameters mixed well and reached stability. Table
3.6 presents the results (posterior mean, posterior standard deviation and 95% confidence limits)
for the model parameters. A Bayesian P-value of 0.50 was calculated, which represents the best
fit model and therefore suggests that this model represents the most parsimonious cheetah
population density estimate. The model estimated 0.61 ± 0.18 adult cheetahs per 100 km² with a
95% maximum of 0.9 and minimum of 0.3 cheetahs/100km². An absolute abundance of 4
cheetahs (range: 2 - 6 individuals) was estimated for the ~700 km² reserve. However, NOTUGRE
45
is probably too small to contain the home ranges of all resident cheetahs and it is therefore likely
that the absolute abundance at any one time may be higher.
Table 3.6 Density estimates of cheetah using MDM buffer width of 28 km and MCMC
parameters set at 200 000 iterations and 4000 burn-in generations. Density is expressed as the
number of cheetahs per 100 square kilometres.
Variables Mean SD 95% Lower HPD level
95% Upper HPD Level
Bayesian posterior probability
sigma 5.10 1.02 3.27 7.14 0.5 lam0 0.03 0.03 0.01 0.07 Psi 0.67 0.20 0.32 1.00
Nsuper 28.56 8.38 14.00 42.00 Density 0.61 0.18 0.30 0.90
3.3.6 SECR analyses for the extended survey
The population closure test was again inconclusive due to insufficient data. The SECR population
estimate for the 130-day period produced a density estimate of 0.58 ± 2.0 adult cheetahs/100 km²
(abundance of 4 cheetahs; range 1 – 6 individuals), using the same MCMC parameters that were
used for the refined model (Table 3.7). This estimate is slightly lower than the density estimated
in the 90 day survey (0.61 ± 0.18 cheetahs/100 km²) (Table 3.6), but the population means did
not differ significantly (t-test; t = 1.22; df =226; p > 0.05). Summaries of the extended survey are
shown in Table 3.7.
Table 3.7 Density estimates calculated from capture histories of the extended survey using the
MDM buffer width of 28 km and MCMC parameters set at 200 000 iterations and 4000 burn-in
generations. Density is expressed as the number of cheetahs per 100 km².
Variables Mean SD 95% Lower HPD level
95% Upper HPD Level
Bayesian P- value
sigma 6.29 1.57 3.78 9.48 0.49 lam0 0.01 0.01 0.00 0.02 Psi 0.64 0.22 0.25 1.00
Nsuper 27.10 9.23 11.00 42.00 Density 0.58 0.20 0.24 0.90
3.4 Discussion
When comparing methods, the level of confidence in the results is usually described as precision,
with high precision referring to relative certainty in the estimation of parameters (Long et al.
46
2008). To have estimates as close to actual numbers as possible (and recognising that these
fluctuate) it is important to minimise bias and this is achieved in the design of the survey (Long
et al. 2008). The design itself relies heavily on the biogeographic characteristics of the species
and the objective of the survey. Estimating abundances and densities of rare species requires
more effort as detection rates will be much lower and this must be taken into account.
Furthermore, certain sampling methods may be effective only for particular species; this is mostly
due to habitat characteristics that strongly influence animal movements and therefore rates of
encounter (Karanth & Nichols 2002; Long et al. 2008). Species that are found in dense bush may
be forced to move along natural trails; so placing camera traps on these well-defined paths
accounts for such habitat heterogeneity (Karanth & Nichols 1998; Henschel & Ray 2003; Long
et al. 2008). However, it may be more difficult to predict movements of species, such as cheetah,
that occur in more open landscapes. Typical camera trapping surveys have camera trap stations
set out systematically across the landscape (i.e. the first survey) and camera traps are placed along
animal trails, however, cheetah capture was low with the only valid captures taken from a camera
trap set at a known scent marking post. The low photo-capture rate of the cheetahs was attributed
to low detectability. Given that cheetahs occur at low population densities (Caro 1994) and the
unpredictable nature of their movements, the location and placement of camera traps is a critical
component to a successful camera trapping survey (Karanth & Nichols 2002, Blake & Mosquera
2014). The design of the survey should have camera traps placed to maximize capture probability
(Karanth & Nichols 2002; Henschel & Ray 2003). The second survey had camera traps located
at cheetah scent-marking posts identified by local guides. Scent-marking posts provided ideal set
up locations as they were frequently utilised by cheetahs. In addition to high probabilities of
cheetah captures, cheetahs would stay at the scent-marking post long enough to obtain clear
photographs; often sniffing the tree and scent-marking for a few minutes before moving on. This
would not only give the camera the chance to capture the subject moving but also often resulted
in multiple photos of an individual during a single capture event, sometimes providing a full
individual profile (i.e. left- and right-hand side photographs). This was also noted by Marnewick
et al. (2006). A possible drawback to such a camera trapping survey is the possible variation in
individual detectability, particularly in relation to age, sex and dominance (Otis et al. 1978). It
has been observed that female cheetahs may use scent-marking posts less frequently than males
and this difference in detection probability may bias estimates and under-estimate population
abundance (Marker 2002; Marnewick et al. 2006; Marker et al. 2008b). Female cheetahs rarely
scent-mark, unless they are in oestrous (Marnewick et al. 2006). However, in the second survey
two females were photo-captured at three scent-marking trees on three different occasions. These
females were not believed to be in oestrous at the time (pers. obs.). Although males use scent
marking posts more frequently than females (Bothma & Walker 1999; Marnewick et al. 2006), 47
provided that sufficient devices are used and the study is carried out over a sufficiently long
survey period, this bias should have a minimal effect on the results. Alternatively, where sample
size permits these sources of heterogeneity can be addressed by including sex-specific encounter
rates but this survey did not have sufficient recaptures to allow such stratification (O’Connell et
al. 2011).
A capture-recapture study requires a relatively large number of recaptures to produce precise
results (Otis et al. 1978; Long et al. 2008). However, sample size is affected by the size of the
sampled area, the number of camera traps used, and the number of trapping occasions and, most
importantly, on capture probability (Otis et al. 1978). Sampling effort can be controlled through
the size of the sampled area and the number of camera traps used (Karanth 1995). It is
traditionally recommended that the surveyed area be at least four times the size of the average
home range of the target species (Otis et al. 1978), but this is logistically and financially
impractical for a wide range of vertebrate species (Foster & Harmsen 2012).
Alternatively, the duration of the survey can be extended judiciously. Ninety days is the
recommended maximum number of days to maintain the population closure assumption when
studying large felids (Karanth & Nichols 2002). However, when surveying for species that occur
at very low population densities, such as cheetahs (Caro 1994) this recommended maximum
number of trapping occasions may be insufficient due to the small sample size and high latency
to first detection which may be due to the cheetah’s large home range. Lengthening the sampling
duration beyond this maximum may improve the robustness of the results but requires careful
consideration of the temporal closure assumption (Foster & Harmsen 2012). Nonetheless,
increasing the length of the survey may be appropriate for some species with long life
expectancies and to areas with long seasons (O’Brien & Kinnaird 2011). Furthermore,
SPACECAP can calculate the density of demographically open populations (Gopalaswamy et al.
2012). Extending the total number of sampling occasions in this study provided substantially
more photographs and independent captures, increasing the degree of certainty in the associated
density (O’Brien & Kinnaird 2011). However, no new individuals were caught and a larger
sample size appeared to change the density estimate little, suggesting that a 90-day survey
provided sufficient data for a robust population estimate. The spatial scale of the study area (±240
km²) may have been insufficient to incorporate sufficient home ranges of cheetahs, thereby
rendering the population closure test inefficient (Otis et al. 1978).
The classical likelihood-based capture-recapture (CR) methods are often preferred due to their
simple formulae and procedures for carrying inference, including calculating standard errors, 48
model selection by Akaike’s Information Criterion and assessing goodness-of-fit (Otis et al.
1978; Karanth & Nichols 1998; Royle et al. 2009a). However, it is difficult to assess the validity
of these procedures particularly when using small sample sizes (Royle et al. 2009a; Gerber et al.
2012). Programs using SECR models such as SPACECAP are believed to be more robust than
conventional CR methods and are thus considered more reliable (Foster & Harmsen 2012;
Gopalaswamy et al. 2012). Among a number of advantages, spatial models allow for
heterogonous detection probabilities among individuals and the Bayesian approach
accommodates for small sample sizes typical of camera trap surveys for low density species
Density estimates calculated using different buffer widths showed considerable differences in
this study. Density estimates will be overestimated if the state-space area is too small. To
overcome this sensitivity, the data were analysed using different buffer widths until the values of
the variables stabilised and models gave an adequate Bayesian P-value. MCMC parameters also
need to be carefully considered to insure that MCMC chains reach stationarity (Noss et al. 2012).
Methods of calculating the buffer width are clearly important when estimating density. The buffer
width used should encompass the maximum movements of individuals caught on camera traps
(Otis et al. 1978; Balme et al. 2009). However, the conventional MMDM method using capture
location data of cheetahs from camera traps was avoided as the surveyed area was likely to be
smaller than the average home range of a cheetah in the study. A cheetah fitted with a satellite
collar (CF3) photographed in this survey had a home range that expanded beyond the entire
survey width (E. Brassine, unpublished data; Figure 3.5). The MDM buffer width was therefore
used to encompass all possible large home ranges. Nonetheless, it should be noted that the
variation in calculated densities may also be an artefact of the small dataset with limited
recaptures, as small sample size and insufficient recaptures are known to compromise the
robustness of the analyses (Otis et al. 1978). Although it is important to have a standardised
approach when designing a camera trapping survey and calculating buffer distances, it is equally
important to understand the limitations of individual surveys and adapt the method to report
population size estimates as accurately as possible. However, the methods and the justification
for their use should be reported in detail so that the study can be adequately repeated and
compared across different study populations.
49
Figure 3.5 95% Kernel UD home range estimates for the collared cheetah (CF3) with GPS fixes used in the home range analysis. The camera trapping surveyed area is also presented.
Although SPACECAP takes into account unsuitable habitats, it is left to authors to interpret this
as they see fit. Hence, in this study, only areas with high human disturbance and large water
bodies were considered unsuitable and excluded. Elevation, vegetation type, prey availability and
competitor avoidance are all likely to affect habitat suitability and hence cheetah densities (Ray
et al. 2005; Pettorelli et al. 2009). For instance, cheetahs avoid lions and an inverse relationship
in their densities has been documented (Durant et al. 2004). Also, prey density outside the
protected area is likely to be low and human encounters (e.g. with poachers) are not necessarily
restricted to human habitations, which may influence the expected density of cheetahs (Pettorelli
et al. 2009). Another drawback to the method is its inability to incorporate captures of
unidentifiable individuals. Nonetheless, the program incorporates aspects of trap location, active
days and capture locations and thus makes it a more powerful tool than the CAPTURE program
using conventional capture-recapture models and the MMDM approach.
3.5 Conclusion
Obtaining reliable population estimates for cheetahs is particularly challenging. This study is the
most intensive study of cheetahs using camera traps and SECR analysis to date. Furthermore, the
results provide the first population estimate for cheetahs in NOTUGRE.
50
This study demonstrates that using species-specific targeted placement and increasing the number
of trapping days for low population densities can produce larger sample sizes for more reliable
density estimates and thereby increasing the precision of the results. Camera trapping for cheetahs
needs to be performed over a large area, over a long survey period and at cheetah-specific
sampling points to calculate robust density estimates. I would recommend integrating multiple
survey methods when assessing population sizes as this can contribute additional information and
cross-validate the quality of the results (Long et al. 2008). In this study, photographs from the
photographic survey (see Chapter 4) were used to construct individual cheetah profiles and aid
in the identification of cheetahs captured in the camera trapping survey. The method can easily
be replicated to perform long-term population monitoring, it is non-invasive and requires
minimum personnel in the field. Finally, density results from within the reserve should not be
used to extrapolate density outside the study area as there are differences in vegetation cover,
prey density, human activity, land-use, and density of other large carnivores.
51
CHAPTER 4 CITIZEN SCIENCE IN CHEETAH RESEARCH: ESTABLISHING POPULATION
ESTIMATES AND SPACE-USE OF CHEETAHS BY WAY OF A PHOTOGRAPHIC
SURVEY
4.1 Introduction
The considerable global population decline and range contraction of the cheetah (Acinonyx
jubatus) is well documented and is attributed predominantly to prey depletion, habitat
degradation and conflict with humans (Marker 2002; Ray et al. 2005; Marker et al. 2010).
Conservation management of cheetahs therefore relies on reliable assessments of the status of
individual populations and the drivers of population trends.
The wealth of information on cheetah ecology, including behaviour, reproduction, ranging
patterns and ecological requirements has primarily been generated from long-term studies in the
Serengeti National Park, Tanzania (Caro 1994; Kelly et al. 1998; Kelly & Durant 2000; Kelly
2001; Durant et al. 2007), and more recently from studies undertaken in Namibia (Marker 1998,
2000; Marker et al. 2003b, 2008a). A key benefit to long-term studies is that they supply vital
information on population trends and demographic parameters of a population, these are
important in understanding population dynamics and viability, information that is crucial for the
conservation management of cheetahs (Durant et al. 2007; Durant et al. 2010; Marnewick &
Davies-Mostert 2012).
Research on the cheetah populations of Botswana has been limited, despite Botswana being
considered a stronghold for the species in southern Africa (Bashir et al. 2004; Ray et al. 2005;
Purchase et al. 2007). Unfortunately, the long-term studies that are required to better understand
the status of cheetahs in Botswana do not exist, and population viability analyses require many
years of research involving significant financial and human resources (Durant et al. 2007).
However, there are a number of alternative techniques for collecting meaningful information on
cheetah population status and distribution over a relatively short period (Bashir et al. 2004).
Cheetahs occur at low population densities and have large home ranges (Caro 1994) so the
probability of locating individuals in the wild is low, making them difficult to survey (Marnewick
& Davies-Mostert 2012). Cheetahs are individually recognisable from their unique spot patterns
(Durant et al. 2007; Kelly et al. 2008) so reliable population parameters, including abundance
52
and demographics, can be determined by reliable identification of individuals. Drawing on
incidental sighting records, sightings and photographs of cheetahs taken by the general public
can be a useful method for monitoring cheetah population trends for well-visited areas and areas
that have habituated cheetahs (Kemp & Mills 2005; Durant et al. 2007; Marnewick & Davies-
Mostert 2012). Photographic survey is a method used for estimating wildlife abundance and
demographics of individually recognisable species by using people already in the field (citizen
scientists). Such an approach reduces labour costs and can be conducted over large spatial scales
and short time spans, which may make estimation of low density species possible where other
methods have proven ineffective (Gros et al. 1996). Furthermore, the method indirectly generates
public awareness of conservation issues, which may create a sense of stewardship (Marnewick
& Davies-Mostert 2012).
Photographic censuses of cheetahs have been conducted in the Kruger National park (KNP),
South Africa (Bowland & Mills 1994; Kemp & Mills 2005; Marnewick & Davies-Mostert 2012;
Marnewick et al. 2014), the Kgalagadi Transfrontier Park, South Africa (Knight 1999) and the
Timbavati Private Nature Reserve adjacent to KNP (Dyer 2013). National Parks have high
numbers of visitors allowing for high search effort (Marnewick et al. 2014), while commercial
private game reserves offer high-end, exclusive game viewing and usually have fewer visitors
than National Parks. Non-commercial private reserves are generally visited even less frequently.
Thus, conducting a photographic survey on a privately owned game reserve may be challenging,
but game viewing on most private reserves is not limited to roads, potentially increasing the
detectability of cheetahs. Furthermore, owners, shareholders, guides and returning visitors may
provide a source of both recent and older photographic records of cheetahs within their reserves;
a wealth of information that would otherwise not have been collected. Historical sightings with
photographic records may therefore be used for temporal comparisons and for estimating age and
family relations of known individuals.
The Northern Tuli Game Reserve (NOTUGRE), in Botswana consists of 36 privately owned
properties; some are used for commercial ecotourism and others on a more exclusive, private
basis. Little is known about the status of cheetahs in NOTUGRE and because cheetahs occur at
low population densities and have large space requirements, they may range beyond its borders
(Durant et al. 2007). Farmlands outside of protected areas may provide refuges from dominant
predators such as lions (Panthera leo) (Purchase et al. 2007; Marker et al. 2010), but increase the
risk of threats from humans such as direct persecution from livestock farmers (Marker 1998).
Areas outside protected reserves, therefore, are important for the survival of cheetahs and the
53
current distribution and movement of cheetahs both within and outside formally protected areas
are hence vital components for the conservation of cheetahs.
In this chapter I explore the suitability and effectiveness of a photographic survey in a private
game reserve to estimate the minimum population size and status of the cheetahs of NOTUGRE.
In addition, the GPS locations of recognisable cheetahs allowed for a preliminary assessment of
cheetah distribution, home range size and the possible movement of cheetahs across international
boundaries.
4.2 Methods
4.2.1 Data collection
Between January 2012 and November 2013, tourists, staff members, shareholders and
neighbouring residents were asked to submit photographs and details (e.g. location, date and time
of sighting, group number and sexes) of any cheetah sightings within or adjacent to NOTUGRE.
At each tourist lodge, experienced field guides took tourists on wildlife-viewing game drives
twice daily, tracking and locating sought-after species, including cheetahs. Digital cameras (n =
4) (either a Canon PowerShot SX260HS or a Nikon Coolpix S9300) with built-in GPSs were
given to field guides who were asked to photograph any cheetahs seen on such drives. In addition,
on most days during the study period, I would drive out into the reserve, actively looking for
cheetahs and following up on reports of sightings. At each sighting, the primary aim was to obtain
clear photographs of the left and right flanks of every individual. As many other photographs as
possible, showing different positions, were also taken to build a complete individual profile for
each cheetah (Maddock & Mills 1994).
Cheetah photographs generated from the camera trapping survey (see Chapter 3) were also used
to supplement the dataset. Moreover a number of residents had camera traps set out for either
recreational or research purposes which captured cheetahs. All cheetah images from camera traps
were included in the dataset.
This photographic survey was promoted through the distribution of pamphlets at all 17 camps
within NOTUGRE (Figure 4.1). All staff and shareholders of the properties were informed of the
survey through email and personal contact. Furthermore, I was based at the largest commercial
lodge (Mashatu Main Camp) and actively promoted and encouraged visitors and staff to submit
their cheetah photographs and sightings information. This approach also enabled visitors to learn
more about the conservation of cheetahs and the cheetah research taking place on the reserve. 54
The survey was advertised on the Mashatu Game Reserve website (www.mashatu.com) and
several popular blogs (e.g. www.blog.mashatu.com). In addition, appeals for cheetah
photographs were made on a local news website, DUMELANG (www.dumelangmusina.co.za),
and appeals were also made at the local Greater Mapungubwe Network meetings (Minutes from
quarterly meetings 2012, 2013).
Figure 4.1 An example of the pamphlet (front and back) distributed to all the camps in NOTUGRE to create awareness of the cheetah photographic census.
4.2.2 Data collation
Digital photographs were received via email and by hand after approaching visitors and reserve
managers at the various lodges. Historical photographic sightings dated back to January 2006 and
included data until November 2013, with the majority of sightings received in 2013.
Each sighting was entered into a Microsoft Excel spreadsheet and included all available
biological information on the sighting (viz. Cheetah ID, group name, time, date, season, number
of individuals, age classes, sexes, location, activity and prey item if seen feeding) and details of
the photographer, including name, contact details and the number of photographs per sighting.
55
Sightings were sorted by sighting date. To avoid autocorrelation of sightings, sightings of the
same individual or group of cheetahs on the same morning or afternoon were pooled into one
sighting event such that a maximum of two sightings per individual or group were recorded per
day (i.e. morning and afternoon). All photographs were stored on an external hard drive
(Transcend StoreJet 1TB) in folders identified by the photographer’s name.
The program Adobe Photoshop Lightroom 5.0 was used to manage all photographic data.
Lightroom is an image data management program that allows the user to organise and catalogue
large numbers of photographs. Furthermore, the program automatically reads the metadata of the
photographs, such as the date and time of capture, and allows for tagging with specific keywords,
making it easier to sort the image database. All photographs were organised into virtual ‘Photo
Collections’ according to each individual cheetah; whereby all photographs of the same
individual were grouped into one virtual folder. The program makes a virtual copy of the
photographs in these folders such that the original photographs are not moved from their original
locations. All photographs were tagged with all relevant information including group size, sex
and the cheetah’s identity number which allowed filtering of these specific criteria and simplified
searching for matches. The software also provides a secondary window display which was used
to compare photographs to aid in identifying individuals.
4.2.3 Individual identification
All photographs were examined by eye and individuals were identified based on unique spot
patterns following the method described in Chapter 3. Each cheetah was assigned a unique
identity number consisting of two letters and a number. The first letter referred to the species (C
= Cheetah), the second letter, the sex (F = Female; M = Male; US = Unknown sex). The number
referred to the position in the identification sequence. Individual profiles were created for each
cheetah; if only one side of the cheetah was available, half profiles were created. A cheetah
identikit of all identified individuals was created for reference purposes. (Appendix IV).
Some sightings could not be used for population estimation because they either did not have
photographic records or were accompanied by photographs from which the cheetah could not be
identified. However, both of these types of sightings could be used for distribution mapping to
display the overall occurrence of cheetahs across the landscape. Some photographs were
submitted with no supporting data. In these instances, the individuals in the photographs were
identified but not used for population estimation or distribution mapping.
56
4.2.4 Long term trends and population demography
Seasons were categorized based on the amount of rainfall distributed over the year on a monthly
basis, using the rainfall records between 1996 and 2013 from three different locations within
NOTUGRE. Two main seasons are experienced, the drought or dry season and the wet season
(Balinsky 1962). The dry season is the period during which there is typically less than 10 mm of
rainfall per month for at least three consecutive months (Balinsky 1962). During the dry season,
trees lose their leaves, grass dries up and becomes exposed. The wet season is defined as the
months where the bulk of the annual rainfall is received (±90% annual rainfall over a 6 month
period). The two seasons are accompanied by changing average maximum and minimum
temperatures. The dry season has lower average minimum and maximum temperatures than the
wet season. Accordingly, the cool dry season was defined as the months of May to October
(monthly rainfall: 7.2mm; monthly maximum temperature: 27.6°C; monthly minimum
temperature: 12.7°C) and the hot wet season was defined as the period from November to April
(monthly rainfall: 59.3mm; monthly maximum temperature: 32.7°C; monthly minimum
temperature: 21.5°C). Seasonal differences in the total number of cheetah sightings were
calculated for the entire survey period (Jan 2006 – Nov 2013).
The collection of photographic records was used to estimate population numbers, minimum
estimated age and family relatedness. To assess general population trends the number of
individuals (regardless of the frequency of sightings) was summed for each year. Individuals that
had probably died were subtracted from this total. The minimum estimated ages for individuals
seen continuously over more than one year were calculated at the end of 2013. The number of
times an individual cheetah or cheetah group was sighted varied. Thus, the minimum estimated
age and family relatedness could only be determined for individuals that had a sufficient number
of repeat sightings and were seen regularly (≥ 5 sightings) for at least a year (n = 16 individuals).
Photographs of cubs provided a means to determine relatedness (Kelly 2001) and approximate
ages of individuals in the population without the use of genealogical records. The approximate
date of birth of a cheetah was estimated as precisely as possible using the method described by
Caro (1994); if the cheetah was first sighted as an adult, it was assumed to have been born at least
two years prior to the first sighting (Caro 1994), therefore estimated ages for adult animals are to
be considered minimum values (Kelly et al. 1998). Cubs were aged by comparing their body
sizes against that of known-aged cheetahs using the aging scale described by Caro (1994). The
first sighting of individuals was used to back-date the approximate date of birth (Caro 1994).
57
4.2.5 Photographic survey - Minimum population size (NOTUGRE)
The minimum population size was estimated following the methods of previous cheetah
photographic censuses conducted in the KNP (Kemp & Mills 2005; Marnewick & Davies-
Mostert 2012, Marnewick et al. 2014). The analysis for the minimum population size for
NOTUGRE included all photographic sightings within NOTUGRE over a seven-month period
(April – October 2013) that was selected because it included the greatest number of sightings
(22.7% of all sightings and 52.2% of all submitted photographs) and overlapped with the camera
trapping survey (see Chapter 3). Thus, all photographs, including those taken by camera traps,
could be included in the analysis. All cheetahs that were realistically alive on the 1st of July 2013
were included in the estimate; this included all cheetahs that had been recorded for the three
months prior to this date (April - June) and all adults and sub-adult (> 3 months old) cheetahs
that were sighted during the four months after this date (July - October) as they would still have
been alive during the census months (Marnewick et al. 2014).
4.2.6 Distribution and home range
All cheetah locations (n = 395) from sightings between 2006 and 2013 were used to asses space
use. Locations were acquired from direct observations using global positioning systems (GPS,
Garmin GPSMAP 62) or by using the closest known landmark if a GPS location was not
available. Some sighting reports were not accompanied with accurate locations; in these cases, if
the name of the property was known, a central location on the property was recorded
(Watermeyer 2012). The location of the sighting was given a rating based on accuracy. A 1 was
given if it was the exact location (within 50 m), a 2 was given if the location was of the closest
prominently-known landmark (within 2 km) and a 3 if the location was inaccurate (> 2 km). The
distribution of all sightings was mapped and kernel utilization distribution (UD) method was used
to represent the areas which reported most sightings.
All geocoordinates were expressed in decimal degrees and imported into ArcMap 10 (ESRI,
Redlands California, USA) for spatial analysis. Geographic distribution was measured as the
extent of occurrence (home range size). The extent of occurrence (Lindsey et al. 2004) was
calculated using the Minimum Convex Polygon (100% MCP) method (Worton 1987). 100%
MCPs are created by joining the outermost location points and the total area within the polygon
represents the individual’s home range size (Worton 1995; Lindsey et al. 2004). Although the
fixed kernel utilization distribution (UD) method is commonly preferred over the MCP technique
as a home range estimator (Worton 1987; Harris et al. 1990; Börger et al. 2006), the MCP 100%
method was more appropriate in this instance due to the opportunistic and probably spatially and 58
temporally biased nature of the data collection and the small sample sizes. The Kernel UD method
calculates home range size of an animal based on the relative amount of time it spends in different
areas of the range (utilization distribution) (Seaman & Powell 1996). The density of points
throughout its range represents the relative amount of time spent in that particular area (Seaman
& Powell 1996). Therefore, density estimates will be high in areas with many location fixes and
low in areas with fewer fixes (Seaman & Powell 1996). In my study, cheetah sightings were
collected on an opportunistic basis, thus the area with greatest observer activity had an increased
chance of cheetah sightings and was inevitably biased in terms of sighting frequency and location.
In addition, the core area driven by game viewing vehicles made the likelihood of sighting
cheetahs in other areas extremely low. Therefore, the density of locations would not necessarily
represent the amount of time an animal spent in a particular area (Seaman & Powell 1996).
Furthermore, areas with no recorded sightings do not necessarily indicate the absence of cheetah
occurrence but rather that no sightings were reported (Lindsey et al. 2004). Absence of sightings
may either mean that there were no cheetahs in that area or that they were present but not
recorded.
100% MCPs (km²) were calculated for all adults (females, including when they had cubs, single
males and male coalitions) that had three or more valid location points (n = 9) (Marnewick &
Davies-Mostert 2012). The locations of cubs were only recorded once they became independent
from their mother and either seen alone or in a sibling group. I investigated the influence of the
number of location fixes on individual/group home range sizes (100% MCP) using regression
analysis (STATISTICA 12). The home range overlap between individuals/groups was
determined by calculating the proportion of area shared with other cheetahs by dividing the total
shared area by the home range size.
4.3 Results
A total of 447 cheetah sightings amounting to 13179 photographs were received from participants
within NOTUGRE (89.0%), properties in South Africa within the Greater Mapungubwe Area
(6.1%), and the Tuli Circle and Sentinel Game Farm area of Zimbabwe (4.9%). Eighty-nine
sightings (19.9%) were received without photographs and 38 sightings (8.5%) did not have
accompanying location data.
4.3.1 Long term trends and population demography
The frequency of sightings varied between seasons, being higher in the cool dry season (n = 264)
than during the hot wet season (n = 179). Four sightings lacked accompanying dates. Thirty-two 59
cheetahs (18 males and 14 females) of all age classes were identified within and adjacent to
NOTUGRE between 2006 and 2013 (Appendix IV; Appendix V). A further 13 individuals of
unknown sex were also identified, but they were only sighted once or twice and only had half
profiles. Age classes at first sightings are shown in Appendix V. To maintain a conservative
overall estimate, only cheetahs with left-side profiles (n = 35; Appendix IV) were included in the
total estimate for NOTUGRE. To assess the general population trend, the total number of
individuals identified (regardless of their frequencies of sighting) in each year was compared
(Figure 4.2). Although fewer photographic and sightings records were received for the years prior
to 2013 (Figure 4.2), the calculated population sizes for 2006, 2010, 2011 and 2012 were higher
than 2013, although only marginally (Figure 4.2).
Figure 4.2 The minimum number of cheetahs identified within NOTUGRE (all age classes), Botswana between 2006 and 2013. The number of sightings per year is also indicated (blue line).
The sighting frequencies, estimated date of birth and relatedness for identified cheetahs are shown
in Appendix V for individual cheetahs identified during the photographic survey period (January
2006-November 2013). The number of times individuals were sighted varied. Nineteen
individuals (12 adults; 7 cubs) only had a single sighting; 26 (full and half profiles) were re-
sighted at least once (> 1 sighting); four (1 adult; 3 cubs) were re-sighted twice, and 22 (10 adults;
12 cubs) were re-sighted three or more times and were therefore considered to be resident
individuals (Kelly 2001). Cheetahs with two or less sightings could have been vagrants (Gros et
al. 1996). Individuals in family groups or stable coalitions that suddenly disappeared were
considered to have died (Balm et al. 2012).
60
A coalition of three males had extensive photographic records, including sightings dating back
to 2006 (Appendix V). Two of the three could be confirmed as brothers (CM1 and CM2) in a
litter of four as they were first photographed when they were approximately 5 – 6 months old.
The same litter was photographed about eight months later with only two cubs remaining (CM1,
CM2) and confirmed by two further sightings a few days later. However, the third individual
(CM3) was not a littermate and joined the two litter-mates at a later stage.
The age of cheetahs was calculated at the end of 2013. CM1 and CM2 were roughly 7.5 years
old. CM3 was at least 7 years old (first photographed as an adult 1 May 2008). CF1, CF2 and
CF3 all appear to be long-term residents and were estimated to be at least 9, 8 and 7 years old
respectively. The other cheetahs with known birth dates were all cubs and sub-adults.
Eight of the 45 (17.8%) cheetahs identified were photographed by camera trapping only,
including a resident coalition of two (CM5 and CM6) with part of their territory stretching over
the central part of NOTUGRE.
4.3.2 Minimum population size (NOTUGRE)
Sightings within NOTUGRE between April and October 2013 amounted to 100 sightings and
6886 photographs. Only two sightings had cheetahs that could not be identified, but another group
member could clearly be identified in both of these sightings. The majority of the sightings (86%)
were from the properties on which Mashatu Game Reserve operates. A total of 13 cheetahs (nine
adults and four cubs) were identified; three more cubs were omitted from the estimate as they
were known to have died (cause of death unknown). Thus, the minimum population size for
NOTUGRE on 1 July 2013 was estimated to be 10 individuals (nine adults and one cub).
Overall, five cheetah social groups were identified that comprised nine adult cheetahs (six males
and three females) and one female cub. Two male cheetahs in a coalition were detected by camera
trapping only, presumably due to their skittish nature. The demographics of the NOTUGRE
cheetah population were: two male coalitions (a coalition of three and a coalition of two), a
sibling group (consisting of one male and one female litter-mate that had recently left their
mother), one lone female, and one family group (consisting of a mother and female cub). The
adult sex ratio over this period was not estimated due to the small sample size. All individuals
accounted for in the minimum population size were believed to be resident as they were all
sighted regularly (Appendix V) and the males were seen scent marking, a behaviour shown by
resident territorial males (Caro 1994).
61
4.3.3 Distribution and home range
The locations of cheetah sightings were unevenly distributed across the reserve. The properties
driven by the tourism operators showed a higher density of cheetah sightings (Figure 4.3). Certain
properties have no commercial lodges and are rarely driven by field guides (Figure 4.3). Thus,
caution needs to be exercised when interpreting the overall distribution of cheetahs, although
higher densities in the reserve are expected due to lower human interference.
Figure 4.3 Map depicting the distribution of all cheetah sightings within and outside NOTUGRE between January 2006 and November 2013, the dark green sections are properties used by commercial tourism operations. 95% and 50% Kernel UD distribution range estimates for the location of sightings are also presented.
Location fixes were predominantly from precise GPS fixes (accuracy 1) (45%) but fixes with a
precision of 2 (33%) and, or 3 (22%) were also used to increase sample sizes. Home range sizes
(MCP 100%) varied markedly across their distribution (Table 4.1). The largest estimated home
range was for female CF3 that used an MCP 100% of 256 km² (Figure 4.4). The second largest
estimate of 188 km² was for the all-male coalition of CM1, CM2 and CM3 (Table 4.1; Figure
4.4).
Table 4.1 The extent of occurrence (MCP100%) and the number of fixes for all identifiable
cheetahs with three or more location fixes. Location fixes were derived from all sightings records
between January 2006 and November 2013, the dates of the first and last locations as well as
timespan are also presented here. Home ranges of cubs were only calculated from time of
independence.
62
Cheetah ID MCP 100% (km²)
Number of fixes
Date of first location
Date of last location
Timespan (days)
CF3 256 134 03/10/2009 15/11/2013 1505 CM1, CM2, CM3 188 72 24/01/2009 28/10/2013 1739
CF1 157 22 17/01/2006 08/01/2013 2549 CF2 155 32 18/11/2008 09/10/2012 1701 CF4 117 15 27/05/2013 22/10/2013 148 CF6 78 7 09/02/2012 12/01/2013 338 CF9 27 3 07/02/2013 30/03/2013 51
CM5 & CM6 24 7 28/06/2013 29/10/2013 124 CF12 0.21 3 06/08/2006 31/01/2007 179
Mean ±SD. 125.3±80.3 36.5±45.2 926±945.3
The centre and south of the reserve are mostly used by cheetahs with many overlapping home
ranges (Figure 4.4). The western corner of the reserve was seemingly utilized only by CF6 with
a little overlap with CF2, CF3 and CF4.
Figure 4.4 The location and extent of occurrence (MCP 100%) of all recognisable adult cheetahs and cheetah groups with three or more location points (n = 9) (ArcGIS 10; projected: Transverse-Mercator, spheroid WGS84, central meridian 29; mapping units: meters).
The effect of sampling bias was investigated by plotting the home range sizes (MCP 100%)
against the number of location fixes (Figure 4.5). Home range sizes were significantly driven by
the variation in the number of location fixes used (Figure 4.5). The greater the number of location
fixes, the larger the home range size. However, the fitted curve suggests that there is an upper
limit to the number of location fixes required to reach an asymptote, at approximately 120
location fixes (Figure 4.5). Nonetheless, the number of location fixes should be obtained from a
wide area and home ranges calculated here must be interpreted with caution. 63
Figure 4.5 A scatterplot representing the relationship between the number of location fixes and home range size of individuals with three or more location fixes.
4.3.4 Home range overlap
The home ranges of the cheetahs overlapped considerably (Table 4.2). The two male coalitions
showed wide home range overlap in both space and time (Figure 4.4); the home range of the male
coalition of two (CM5; CM6) fell entirely within the home range of the male coalition of three
(CM1; CM2; CM3). Independent cubs showed similar ranging patterns to their mother’s.
Independent female cubs CF4 and CF9 shared, respectively, 84.6% and 88.9% of their home
ranges with their mother (CF3). The coalition of three males (CM1, CM2, and CM3) shared
61.2% of their home range with their mother (CF1). The home ranges of adult females overlapped
between 0% and 100% (37%). Overlap of a female’s home range with that of a male ranged
between 0% and 100% (49%). The high home range overlap may be an artefact of the small
sample sizes used to calculate some of the home ranges. Additionally, the home ranges were
worked out using all available data thus includes a sampling period spanning over seven years
when ranges are likely to shift.
Table 4.2 The percentage overlap for each adult cheetah’s 100% MCP home range for the
Northern Tuli Game Reserve for individuals alive over the same time period (January 2006 –
November 2013).
Cheetah ID Overlap CF1 CF2 CF3 CF4 CF6 CF9 Coalition
2 Coalition
3 CF1 - 42% 55% 28% 0% 6% 15% 99% CF2 43% - 86% 75% 17% 6% 100% 44% CF3 33% 52% - 39% 18% 9% 9% 50% CF4 38% 100% 85% - 21% 0% 19% 34% CF6 0% 35% 58% 31% - 0% 0% 0% CF9 37% 37% 89% 1% 0% - 4% 67%
Coalition 2 96% 100% 100% 92% 0% 4% - 100% 64
Coalition 3 61% 36% 68% 21% 0% 10% 13% -
4.3.5 Cross boundary movement
Cross boundary movement was observed only between NOTUGRE and the Tuli Circle of
Zimbabwe. Although cheetahs were sighted in South Africa (n = 9 individuals), there was no
evidence for movement across this border. Furthermore, there were no sightings of cheetah within
the Mapungubwe National Park (Cilliers, Section Ranger in Mapungubwe National Parks, pers.
comm.) or the Maramani Communal lands east of NOTUGRE, in Zimbabwe. Cheetah sightings
west of the reserve were also scarce (see Chapter 5).
4.4 Discussion
The photographic survey method was a useful technique to get rapid and seemingly robust
population estimates for cheetahs within NOTUGRE. Public response was excellent, particularly
in the properties which had the most intensive awareness. Personal communication was by far
the best technique for obtaining photographic sightings and the aid of volunteers to assist with
public liaison at various tourist camps should therefore be considered for future censuses. Digital
cameras supplied to field guides were also an efficient tool as photographs were accompanied by
GPS locations and correct dates and times. However, it is important to ensure that the guides
understand the project. Providing rewards for their assistance may also facilitate their support.
Tour guides, including professional photographers, are frequent visitors of game reserves and
were the most useful source of historical cheetah sightings; they also showed enthusiasm and
were eager to assist.
Other indirect methods such as questionnaire surveys are commonly used to provide information
on cheetah populations. Questionnaires that rely solely on sightings information suffer from
inaccurate reporting, including errors in aging and sexing of individuals (Gros et al. 1996).
Furthermore, the method may underestimate population size because all sightings of the same
group compositions are considered to be the same as individuals cannot be distinguished (Gros
et al. 1996). Photographic surveys allow for individual recognition, therefore providing a much
more robust method for population estimates.
Possible drawbacks to photographic surveys include uneven search effort, inaccurate reports,
imprecise metadata and, despite the use of database management software, the time required to
compare and match large quantities of photographs. Regardless of these disadvantages, this
65
method provided important baseline data on the minimum population size, home ranges, ages,
family relations and cross-boundary movement that would otherwise have required years of
intensive research. Furthermore, once a relatively comprehensive cheetah photographic database
exists (such as Appendix IV), this method can be used as a tool to monitor future cheetah
population survival rates and recruitment.
Cheetah sightings were more frequent in the cool dry season. This is possibly because of a more
open landscape and better visibility. In addition, the cool dry season also coincides with the peak
tourism season for photographic destinations. Based on the results of this study it is recommended
that future photographic surveys be carried out over the cool dry season.
4.4.1 Camera traps as an additional tool
Photographic surveys rely on photographs taken opportunistically, hence they are biased to
cheetahs that are relatively habituated and which frequent the most visited areas (Gros et al.
1996). The method also relies on good quality photographs for identification of individuals (Gros
et al. 1996). Most populations will have a few shy individuals that may not be documented at all
(Maddock & Mills 1994; Gros et al. 1996). Forty-seven additional sightings with a total of 517
photographs were captured by camera traps (camera traps set out for research and recreational
purposes) in my study. Camera traps photographed eight otherwise unrecorded individuals and
89% of images from South African properties were taken by camera traps. This suggests that not
all individuals within a population will be detected if only the traditional photographic
submissions are used. Therefore, photographic surveys should not rely solely on the use of
photographs but also supplement the database with camera trapping data.
4.4.2 Minimum population size and trends
The total population sizes calculated over the last eight years (Figure 4.2) could suggest a
decreasing population; however there is insufficient data to draw a conclusive interpretation. It
is expected that when more effort and hence more photographic sightings are received, more
individuals in the population are likely to be detected. However, despite the higher effort,
including awareness, camera trapping and cameras supplied to field guides, the minimum
population size for 2013 was below that of previous years. In addition, camera trapping sightings
were only received for 2012 and 2013 thus shy individuals may have been excluded in the
minimum count of preceding years. Possible reasons for the observed decline may be increased
intra-guild predation with a recovering lion population (Snyman et al. 2014). A high lion density
can have a negative impact on cheetah numbers, particularly if the cheetah population is
fragmented (Laurenson 1994; Kelly et al. 1998; Kelly & Durant 2000; Durant et al. 2004). 66
Human-related mortalities and the persecution of cheetahs outside the boundaries of the reserve
(O.R. Masupe, Community Liaison officer/Anti-poaching officer, pers. comm.) may also be a
cause for their observed decline. During the eight year period, a number of cases of predator
persecution were reported, including the death of a sub-adult cheetah reported from the
community-owned farmland adjacent to NOTUGRE. Furthermore, wire snaring inside the
reserve is a growing concern (P. Le Roux, general manager of Mashatu Game Reserve, pers.
comm.). Although snares are mostly set to capture small to medium antelope for bushmeat,
incidental capture of cheetahs have been recorded in NOTUGRE.
4.4.3 Family relations
Relationships between certain individuals could be determined from photographs of family
groups. The coalition of three males is composed of two littermates and a non-relative. Caro
(1994) observed similar trends and found that most coalitions with three members consisted of
two littermates and a non-relative, whereas most coalitions of two are composed of littermates,
with an average 29.4% of cheetah coalitions containing non-relatives (Caro 1994). More recently
a genetic study conducted in Botswana found that three out of four male coalitions were
genetically unrelated (Dalton et al. 2013).
4.4.4 Estimated age
The minimum estimated ages of the male cheetahs in my study are comparable with that of the
estimated minimum ages of males in the Serengeti (6.0 to 8.4 years = 6.9, SE = 0.2) with
territorial males expected to have a longer lifespan than non-resident males (Caro 1994). Adult
female cheetahs in my study showed a higher lifespan (average 8 years) than that of Serengeti
females (average 6.2 years; Caro 1994) but Serengeti cheetahs were also observed to reach 14
years 5 months (female) and 11 years 10 months (male) (Durant et al. 2010). Furthermore, the
estimated ages for the adult female cheetahs in my study are probably underestimated since these
cheetahs were assigned a minimum age of two when first sighted as adults.
4.4.5 Distribution and home ranges
The observed distribution of sightings identifies areas of cheetah occupancy/presence both within
and outside the formally protected areas. Although the widespread distribution of sightings may
include outliers such as temporal trips outside home ranges and dispersing individuals (Lindsey
et al. 2004), my findings highlight the need for conservation action to include both protected and
unprotected areas.
67
Most of the recorded sightings occurred on commercial properties where ecotourism is the
primary land use. This highlights the effect of sampling bias as more frequent activity on these
properties is evident and guides are actively searching for this species whilst on game viewing
drives. Mashatu Game Reserve is a commercial photo-tourism operation, with two commercial
and five private camps, and a maximum of 15 vehicles operating at any one time. Vehicles used
by researchers, horse safaris, walking safaris, photographic safaris and eco-training may also
operate on the reserve, and certain areas are driven more frequently than others, particularly the
alluvial plains surrounding the major river systems, because of the higher abundance of game in
general and carnivores in particular, therefore more photos are likely to be recorded in these areas
(Figure 4.3).
Nonetheless, cheetahs have large home ranges relative to property sizes and individuals that may
occur on properties seldom driven are likely to be sighted on neighbouring properties and were
thus probably included in the minimum population estimate (Maddock & Mills 1994).
Importantly, only a single sighting of an individual is required to include it in population
estimates.
Home ranges should be interpreted with care as the MCP 100% method provides only a
rudimentary, unsophisticated estimate of home range. Previously reported home ranges varied
between 11 km² and 1651 km² (Pettifer 1981; Caro & Collins 1987; Mills 1989; Caro 1994; Gros
et al. 1996; Hunter 1998; Marker 2000; Purchase & du Toit 2000; Broomhall et al. 2003; Bissett
& Bernard 2006; Bissett 2007). Additionally, Bissett & Bernard (2006) found home range sizes
to differ with different social groups and Bissett (2007) reports that home ranges of female
cheetahs differ with reproductive status. In my study, the home range size for the coalition of
three males (188 km²) was considerably larger than that of the average territory for male cheetahs
in the Serengeti Plains (37.4 km²: Caro 1994). However, the largest home range size I recorded
was used by a female (256 km²) and this is smaller than the smallest female range recorded for a
Serengeti cheetah (394.5 km²: Caro 1994) but larger than home ranges from cheetahs in the KNP
(102 km² to 192 km²: Broomhall et al. 2003). Nonetheless, a very strong correlation was observed
between the number of fixes and home range size (Figure 4.5). Thus, the number of fixes used
may have been insufficient to accurately calculate some individuals’ home ranges.
4.4.6 Home range overlap
Both male coalitions appeared to be resident and territorial as they were repeatedly seen in the
same area and scent-marking at specific scent marking posts (Caro 1994). Furthermore, a scent-
marking post was used by members of both coalitions. Extensive home range overlap in males 68
has been noted in non-resident males in the Serengeti (Caro 1994), but it is not common in
resident males that are territorial.
Individual cheetahs with known family relationships showed considerable spatial overlap,
consistent with previous findings where sibling groups that had recently left their mother were
recorded to have an average of 79.5% of their home range falling within their natal range (Caro
1994). However, the home range of CM1 and CM2 (7.5 years old) overlapped considerably with
that of their mother, implying that they did not disperse from their natal home ranges. This is
unusual for male cheetahs and inbreeding may potentially be a concern (Kelly et al. 1998).
Differences in home range overlap for family and non-family members could not be assessed
because of unknown relatedness between some individuals.
4.4.7 Sightings outside of NOTUGRE
Sightings outside NOTUGRE were scarce (11% of all reports) and reports were mostly from
camera traps, suggesting that although cheetahs occur outside NOTUGRE, they are rare and
skittish. However, restricted access to certain properties outside NOTUGRE prevented a
thorough census and data were limited to reports from a few respondents (n = 6). Very few
sightings in this study were reported from communally-owned farmlands that occur both east and
west of the reserve (Zimbabwe and Botswana). Questionnaire surveys were undertaken in the
communities west of the reserve (see Chapter 5), and few respondents (n = 24) had seen cheetahs
and many farmers could not distinguish between cheetahs and leopards (Panthera pardus).
Scarce cheetah sightings could be a result of poor observer activity or the presence of skittish
individuals, or they could be attributed to depleted prey populations (see Chapter 5). However, it
has been shown that human disturbance alone can limit cheetah occurrence by the alteration of
natural habitat, the presence of domestic dogs (Canis lupus familiaris), bushmeat poaching and
direct persecution (Andresen et al. 2014). Significantly, low prey availability is typically
associated with transient carnivore populations (Winterbach et al. 2014). Habitat fragmentation
and persecution by livestock farmers may further discourage the movement of cheetahs outside
of protected areas (Lindsey et al. 2004).
For a population to have long-term demographic viability, it should comprise over 300
individuals (including cubs), with isolated populations at a higher risk of extinction (Durant et
al. 2007). Most African protected areas are not large enough to contain viable populations of
cheetahs and hence rely on the relocation of cheetahs outside of protected areas, either naturally
or artificially (translocations) (Durant et al. 2007; Macdonald et al. 2010b). Thus, protected areas
such as NOTUGRE are unlikely to support a genetically viable population of cheetahs and must 69
therefore rely on cheetahs persisting beyond the borders of the reserve and immigrating from
time to time. The conservation of cheetahs relies on their survival across extensive areas of
connected habitats with heterogeneous populations of prey and predators (Durant et al. 2007). It
is well documented that cheetahs are adversely affected by human activity and high human
densities can provide an immediate barrier to the movement of cheetahs (Lindsey et al. 2004;
Woodroffe 2000). Persecution, fences and habitat modification may limit the distribution and
dispersal of cheetahs outside NOTUGRE, possibly fragmenting populations into discontinuous
sub-populations. Dispersal and therefore inter-patch connectivity is essential for populations to
persist (Durant et al. 2007; Elliot et al. 2014).
The results of my study suggest that there is little or no exchange with adjacent farmlands in
South Africa. Furthermore, human settlements surround much of the western, southern and
eastern boundaries of the reserve. Reducing direct persecution and mortality is an important focus
for conservation efforts, as well as maintaining habitat connectivity and the wild prey
populations. Establishing a substantial database of sightings and distribution information
provides a better understanding of which areas are important for conservation and brings clarity
to the conservation requirements for cheetahs in this part of Botswana.
4.5 Conclusion
Cheetah population demographics based on the recognition of individuals can provide useful,
quick and relatively inexpensive population estimates. The findings of this study provide baseline
data on the status of cheetahs of NOTUGRE and contribute to a better understanding of cheetah
ecology for a population occurring in an open system that experiences different pressures,
including human activity. Furthermore, the study demonstrates the feasibility of a photographic
survey combined with camera trapping to survey a cheetah population. The use of photographic
records, including recent and old photographs, may provide an alternative to intensive field
studies that involve high financial and time expenses. Comparing recent and old photographs not
only provides current estimates of minimum population size but can also provide information on
changes in population sizes, relatedness and age of frequently seen individuals. Asking the public
for photographs also raises community awareness of these animals.
Future research should focus on the monitoring of cheetahs outside of NOTUGRE to address the
lack of information on the distribution and status of the population outside of the reserve,
including possible connectivity to the NOTUGRE cheetahs. It is further recommended to have
70
on-going population monitoring to understand the drivers of this population, particularly in
relation to human persecution and changes in the numbers of other carnivores.
71
CHAPTER 5 HUMAN-PREDATOR CONFLICT IN AND AROUND THE NORTHERN TULI
GAME RESERVE, BOTSWANA
5.1 Introduction
Human-wildlife conflict (HWC) is a global concern that can arise wherever humans and wildlife
come into contact (Schiess-Meier et al. 2007). Both humans and wildlife are generally negatively
affected; wildlife can threaten economic resources and human lives and the perceived or actual
threat posed by wild animals often resu lts in their persecution (Marker et al. 2003a, 2003c;
Clarke 2012). The worldwide increase in the human population is causing habitat loss and
fragmentation (Ray et al. 2005). As human populations move into previously uninhabited areas,
the potential for conflict between humans and wildlife increases as both humans and wildlife are
forced to compete for the same limited resources (Graham et al. 2005; Clarke 2012). Importantly,
conflict contributes to the decline of wildlife populations, particularly large predators, threatening
the survival of species on a local, regional and, in some instances, global scale (Ogada et al.
2003).
The decline in wildlife populations as a result of persecution has led many countries to
promulgate laws for the protection of the wildlife involved (Graham et al. 2005; Sifuna 2010).
However, when wildlife, and predators in particular, are legally protected, farming communities
often develop resentment towards authorities and wildlife conservation programs, which are
often perceived as unsympathetic towards the farmers’ needs (Mishra et al. 2003; Clarke 2012).
Thus, mitigating HWC requires a thorough understanding of a complex situation and effective
measures will differ from one locality to the next (Dickman 2010; Clarke 2012). Nonetheless,
large carnivore populations can persist and co-exist in human-dominated landscapes where
appropriate wildlife management is established and enforced (Linnell et al. 2001).
Human-predator conflict, particularly over the depredation of livestock, is one of the most
prominent forms of HWC (Ogada et al. 2003; Graham et al. 2005). The occurrence of predators
in human-inhabited areas can lead to conflict due to the perceived threat to human lives (Graham
et al. 2005) but more frequently because of the high financial costs associated with livestock
depredation (Patterson et al. 2004; Graham et al. 2005; Winterbach et al. 2014). Depredation of
livestock may result in revenge killing of predators by farmers (Sifuna 2010; Chattha et al. 2013).
However, persecution may also be indiscriminate and is not always retaliatory, but rather used as
72
a means of prevention (Marker et al. 2003c). The killing of predators is often considered to be
the most cost-effective and efficient way to reduce levels of predation on livestock (Thorn et al.
2014). However, Graham et al. (2005) found that predator density is not related to livestock
depredation, but may rather be a function of prey availability, hence reducing predator
abundances is unlikely to resolve the conflict. Indiscriminate persecution of predators can have
a severe impact on the conservation of threatened species (Ogada et al. 2003; Thorn et al. 2014).
Furthermore, ecosystems are complex, with multi-trophic interactions, therefore removing large
predators from the system can trigger undesirable ecological responses (Graham et al. 2005). For
instance, the absence of large predators on farmlands may trigger ‘meso-predator release’ which
may further exacerbate the livestock-predator conflict (Prugh et al. 2009). Meso-predator release
is the dramatic increase in the abundance of smaller predators which is commonly associated
with the absence of large predators (Prugh et al. 2009).
Large predators typically occur at low densities due to their large space and energy requirements
(Patterson et al. 2004; Ray et al. 2005). Protected areas may therefore be too small to support
viable populations of large predators and connectivity of isolated populations across agricultural
farmlands may be essential for predator survival (Ogada et al. 2003; Selebatso et al. 2008;
Winterbach et al. 2014). Furthermore, subordinate predator species such as cheetahs (Acinonyx
jubatus) and African wild dogs (Lycaon pictus) may actively seek refuge outside protected areas
from more dominant competitors which generally occur at higher population densities within
reserve boundaries (Laurenson et al. 1995; Marker 1998; Marnewick & Cilliers 2006; Klein
2007; Winterbach et al. 2014). For example, about half of Botswana’s cheetah population is
found on farmlands and African wild dogs are widespread across agricultural land (Winterbach
et al. 2014). Human-dominated areas outside protected areas can therefore have a direct impact
on the survival of these wide-ranging carnivores, and may contribute to localized extinctions
(Woodroffe & Ginsberg 1998; Klein 2007; Winterbach et al. 2014).
The protection of large predators outside protected areas is essential and conservation action
plans need to account for this (Laurenson et al. 1995; Woodroffe & Ginsberg 1998; Kelly &
Durant 2000). Reducing predator persecution is crucial in addressing human-predator conflict
(Thorn et al. 2012). Livestock depredation can be prevented by implementing appropriate
livestock husbandry techniques and developing a better understanding of the carnivore species
involved (Ogada et al. 2003). However, mitigating the wildlife damage, such as reducing
livestock loss, alone is unlikely to resolve the conflict as social factors and attitudes towards
predators are also important determinants of human behaviour (Dickman 2010). Importantly,
conflict mitigation can be achieved through a change in attitude towards wildlife, encouraging
73
cooperation and accounting for the concerns of the community, as successful conservation cannot
succeed without the support of local communities that coexist with the wildlife (Sifuna 2010).
Attitude can be defined as ‘a learned predisposition to respond to an object or class of objects in
a consistently favourable or unfavourable way’ (Foddy 1994) and a person’s attitude will directly
affect their choice of action (Hudenko 2012). Thorn et al. (2014) found that farmers who killed
predators had significantly more negative attitudes towards predators than farmers who did not
kill predators. However, livestock losses to predators were not a strong determinant of farmer
attitude (Thorn et al. 2014). Attitudes are mostly formed through evaluations of personal values
and by direct knowledge, but also by emotional responses based on previous experiences and,
most importantly, by social influences (i.e. discussions with neighbours) (Foddy 1994; Thorn et
al. 2012). A positive attitude within and among farming communities may increase tolerance
towards predators. Thus, a sound understanding of the elements which may form and change
attitudes towards predators is essential for conservation planning (Thorn et al. 2014).
Socio-demographic variables such as age, gender, annual income and the level of education may
impact attitudes (Røskaft et al. 2007). A more positive attitude towards predators is generally
observed where people have a better understanding of the environment and its functioning
(Røskaft et al. 2007). Marker et al. (2003a) found that Namibian farmers were less likely to
persecute predators on their farms after participating in an education program about predators.
But, providing information alone is unlikely to change attitudes as humans rarely make rational
decisions; decision making is primarily driven by an individual’s emotional response rather than
by logic and reasoning (Hudenko 2012). However, generating affection and evoking positive
emotions from positive experiences with wildlife can change attitudes (Hudenko 2012; Heberlein
2012). Understanding the drivers of attitude, personal values, and the emotional relationships
which people have with wildlife is therefore important for any conflict mitigation strategy.
Attitudes towards wildlife can be much more positive where conservation efforts include the
welfare of both the wildlife and the people (Sifuna 2010). In Botswana, the Department of
Wildlife and National Parks (DWNP) is responsible for a state funded compensation scheme for
livestock depredation by certain wild predator species (Schiess-Meier et al. 2007). This incentive
was introduced in an attempt to increase tolerance towards predators and reduce HWC. Farmers
are required to report losses of livestock to receive compensation; such claims are then
investigated by a DWNP Problem Animal Control (PAC) officer to ensure that the damage was
caused by a compensated species (Schiess-Meier et al. 2007). Predator species that are considered
74
dangerous (i.e. which cannot safely be chased off) and threatened species such as the cheetahs
and wild dogs are included on this list (Kgathi et al. 2012).
Botswana has a human population of about 2 million (annual growth rate of 1.9%) (Winterbach
et al. 2014), of which about half live in rural villages and small homesteads (known as cattleposts)
on tribal farmland. In 2008, Botswana had a livestock population of approximately 4.5 million
with 92% found on rural communal farmlands (Winterbach et al. 2014). In 2012, the cattle (Bos
primigenius) population was estimated to be over 3.1 million and the small stock (goats Capra
aegagrus hircus and sheep Ovis aries) population was estimated to be 1.6 million (DWNP 2012).
Livestock pastoralism clearly forms an important part of rural economic income and cattle have
a significant cultural value by representing wealth and social standing for local people (Clarke
2012; Winterbach et al. 2014). However, overgrazing by cattle in rural farmlands has led to bush-
encroachment, the growth of unpalatable grasses and increased proportions of bare ground, all of
which will result in less available forage impacting not only the livestock grown but also the wild
prey base (Myers 1975; Klein 2007). Furthermore, due to a growing human population and
increased livestock numbers, farmers and their livestock are encroaching onto the edges of (and
into) protected areas.
Communal farmland is situated along the western border of the Northern Tuli Game Reserve
(NOTUGRE) and farmers regularly encounter large predators, including cheetahs, brown hyenas
(Hyaena brunnea), African wild dogs, lions (Panthera leo), leopards (Panthera pardus), spotted
hyenas (Crocuta crocuta), and black-backed jackals (Canis mesomelas). Depredation of
livestock by these predators occurs in these communities and incidents of revenge killing and
predator persecution by local farmers have been reported (O.R. Masupe, Community Liaison
officer/Anti-poaching officer, pers. comm.). Human-induced mortality is also brought about by
the accidental snaring of predators (Ogada et al. 2003). Wire snares are set both outside and inside
NOTUGRE by poachers for the bushmeat trade (O.R. Masupe pers. comm.).
This chapter investigates human-predator conflict within the communal farmlands bordering onto
and within NOTUGRE, and how it influences farmers’ attitudes towards predators. Specifically,
I assessed the extent of livestock depredation, the nature of depredation events, and how such
conflict influenced attitudes towards the conservation of predators outside protected areas. I also
assessed the relationships between socio-demographic characteristics (i.e. age, education level,
primary source of income, position held on cattle post), livestock husbandry techniques,
knowledge of local predators, livestock losses, and attitudes towards predators. I predicted that a
more positive attitude would be observed among respondents who had a higher level of education 75
and knowledge of predators, but that attitude would be negatively correlated with the extent of
livestock loss (Bath 1998; Ericsson & Heberlein 2003; Røskaft et al. 2007). Respondents who
relied on livestock as a primary form of income would be expected to have a lower tolerance of
livestock loss and would therefore likely have more negative attitudes (Røskaft et al. 2007).
5.2 Methods
5.2.1 Study area
Communal grazing of livestock by local subsistence farmers is the predominant land use on the
tribal land immediately west of NOTUGRE. The boundary between the reserve and the farmland
is formed by an electrified game fence (height: 2.1m with three electrical stands at 1.8m, 50cm,
20cm) which is frequently damaged by elephants (Loxodonta africana) and other wildlife and
therefore does not normally restrict the movements of large carnivores and/or livestock.
Cattleposts (human settlements that include a few individual dwellings and livestock enclosures)
are irregularly spaced across the landscape with an approximate human population density of 381
people/100km² (Klein 2007) and a cattlepost density of 21.2/100km² calculated from GPS
locations acquired during the survey (Figure 5.1). A small village, Lentswe Le Moriti, is situated
within the NOTUGRE boundary and the surrounding land is used for agro-pastoral farming with
a total of 11 cattleposts which have livestock.
Figure 5.1 Map of the study area including the locations of all cattleposts (n = 80) surveyed along the western boundary of and within NOTUGRE (ArcGIS 10; projected: Transverse- Mercator, spheroid WGS84, central meridian 29; map units: meters).
76
The habitat in the communal farmland is generally much more open (in the horizontal plane) than
in NOTUGRE as a result of over-grazing, tree felling and bush clearing (Figure 5.2; Figure 5.3).
Although wildlife is present on the communal farmlands, numbers are generally low due to
habitat degradation and poaching. The main occupation of residents living on cattleposts is
subsistence farming and livestock pastoralism. Livestock kept includes goats, sheep and cattle.
Donkeys (Equus africanus asinus) and horses (Equus ferus caballus) are also kept for transport
purposes and poultry (Gallus gallus domesticus) is important for home consumption. Livestock
is mostly left unprotected during the day, but is brought back into kraals (the traditional name for
livestock enclosures, pens or corrals made up of wooden posts and/or branches) at night.
However, stray animals often sleep out in the field unprotected.
Figure 5.2 Photographs of typical acacia veld inside (top) and outside (bottom) NOTUGRE. The overall habitat inside the reserve is clearly denser with few tall trees; outside the reserve it is typically more open with little undergrowth and a distinct browse line. Photos: Eleanor Brassine.
77
Figure 5.3 Mopane bushveld is typically denser and has stunted growth inside the reserve (top), whereas outside the reserve the habitat is typically more open with taller trees (bottom) as a result of bush clearing and overgrazing. The absence or infrequent occurrence of elephants on farmlands also contributes to taller mopane trees which are mostly absent in the reserve. Photos: Eleanor Brassine.
5.2.2 Data collection
Information to understand the extent and potential drivers of human-predator conflict in the
communities along the western boundary of, and within, NOTUGRE was collected by means of
an interview-based questionnaire (Appendix VI). The questionnaire consisted of 88 questions
and was divided into five sections: demographic and socio-economic characteristics (n = 8
questions); farm details and management practices (n = 36 questions); details of wildlife and
predators in the area (n = 14 questions); predation and conflict (n = 25 questions); perceptions
78
and attitudes towards predators (n = 5 questions). The questionnaire was written in English but
interviews were conducted by translators in the local language, Setswana, where necessary. Two
teams conducted the interviews simultaneously, with each team consisting of a trained researcher
and a local Motswana translator who had a good understanding of local traditions and farming
practices. The questionnaire survey was carried out with the authorisation of the Rhodes
University Ethical Standards Committee (Ethics clearance number: ZOOL-03-2012). The
questionnaire was explained and read over with the assistants prior to the start of the survey. No
pilot study was conducted due to the limited number of cattleposts (n = 80) in the area, but the
questionnaire was adapted from a similar study done in the west of Botswana by Cheetah
Conservation Botswana (CCB) (Klein 2013). Furthermore, the questionnaire was reviewed by
experts in the field from Rhodes University prior to being conducted (Olson 2010).
The study was conducted in the cool dry season between May and August 2012. To ensure even
coverage, and to have as large a sample size as possible, all cattleposts situated within 14 km of
the border of NOTUGRE were surveyed opportunistically (Figure 5.1). A previous study found
that HWC occurred most frequently in areas immediately adjacent to protected areas, with the
highest incidents of damage within five kilometres and up to about 20 kilometres from wildlife
areas (Sifuna 2010). Cattleposts were located by asking for directions from local residents. The
interviews were conducted at cattleposts within the following farming areas: Lentswe Le Moriti,
Fairfields, Mathlabaneng, Sethoba, Malopeng, Letswerang, Motswereng, Monyemotobo,
Lekono, Makadibeng, Semphane, Thune, Matshekge, Madiope, Thebele, Manyehome, and
Mokalati. When occupants of a cattlepost were absent, we would return to it on the following
visit (approximate time between visits was 1 week).
All respondents were interviewed at their cattleposts and the date, time and location (GPS,
Garmin GPSMAP 62) were recorded for each interview. Visiting the cattleposts themselves
allowed for an improved understanding of current farming practices and the methods used to
protect livestock from predators, including the designs of kraals and the distance of kraals from
homesteads. Once the interview was complete, we also took the opportunity to advise on
livestock loss prevention measures, how to identify the more common predators and the status of
cheetahs and their conservation.
Upon arrival at a cattlepost, we would introduce ourselves and explain the nature of the research.
Residents were asked if they wanted to participate in the study, explaining that they had the right
to refuse being interviewed. All of the cattlepost residents we visited (n = 80) agreed to take part
in the questionnaire survey, with each interview lasting an average of 45 minutes. Before 79
commencement of the interview, two posters depicting the common large predators were
presented as supplementary material (Appendix VII) to ensure respondents could identify
common predator species correctly and to assess knowledge of the common predators (Gros
2002). The posters consisted of photographs of eight common large predators, lion, leopard,
cheetah, spotted hyena, brown hyena, African wild dog, blackbacked jackal and domestic dog
(Canis lupus familiaris) (see Appendix VII). The questionnaire had both closed and open-ended
questions, respondents could also provide additional comments if they wished. All answers to
open-ended questions and any comments were recorded in full but later classified into groups
according to their similarities to facilitate statistical analyses (Foddy 1994). Classification may
be subject to a degree of interpretation, but a standard approach was applied when classifying
responses by the principal researcher. This involved reading over and sorting all the responses
into relevant categories (Foddy 1994). The respondents were made aware that they could respond
with “I don’t know” to any question (White et al. 2005). Furthermore, to assess the accuracy of
responses, ground-truthing questions (n = 15) were included which cross-referenced respondent
answers (White et al. 2005). For example, respondents were asked to explain how their livestock
was cared for during the day and night (Appendix VI, section D: questions 17 – 18). Later in the
questionnaire (Appendix VI, section H; questions 63- 72), respondents were asked to provide a
‘yes’ or ‘no’ answer for the livestock husbandry techniques which they were using to protect their
livestock from predators. If their answers differed, we would ask respondents to clarify their
previous answer, and in the instances where the responses remained inconsistent, their
questionnaire was disregarded and removed from all analyses.
Respondents were asked to give details on the number and type of livestock owned, they were
then asked if they suffered any losses to predators, and if so, to give a detailed explanation on
each depredation incident over the previous 18 months. A period of 18 months was chosen as it
coincided to the beginning of 2011 and was thus easier to explain to respondents. Each respondent
was asked to provide the estimated value of livestock in Botswana Pula (BWP), and the average
value for each livestock type was calculated from all responses to calculate the total value of
livestock lost per cattlepost.
Respondents were asked to rank the significance of potential problems faced as a livestock farmer
on a three-point Likert scale (0 – 3). A maximum of three was given for major problems and zero
was given if it was not a problem at all, thus more than one cause could be ranked as a major
cause of livestock loss. The importance of each possible problem was evaluated by summing the
number of maximum values for each cause. We then asked respondents to rank problem predators
and to name any other predator species that they identified as damage causing species. 80
We asked respondents to provide details on wildlife and predator species occurring in the area
and whether they perceived any changes in abundances over the last 10 years. Specific details on
cheetah sightings were also requested, including the number of animals seen, the location and
date. Sightings data were used to assess the occurrence of cheetahs outside the reserve and,
depending on the quality of the responses, to crudely estimate abundance (see Chapter 4).
Cheetah sighting details were only used if respondents could clearly identify cheetahs from the
poster and could provide additional description on the behaviour of cheetahs.
The location of each cattlepost was mapped (Figure 5.1) and distances (km) to the nearest fence
line boundary were calculated using ArcMap 10.0 (ESRI, Redlands, California, USA).
Cattleposts found within the reserve were assigned a distance of 0 km.
5.2.3 Conflict of responses
Although the purpose of the research was clearly stated upon arriving at the cattleposts, I wore
the Mashatu Game Reserve uniform and drove a Mashatu vehicle with Mashatu Game Reserve
and NOTUGRE research stickers clearly visible. Respondent attitude and the answers given may
have therefore been influenced by the fact that I was clearly associated with the reserve. While
some respondents were unhappy with the reserve, others were grateful to see that the reserve was
concerned about their problems with livestock depredation. In an attempt to counteract these
problems, I would retain a neutral position, remaining objective and detached to encourage
respondents to provide an honest response to my questions when conducting interviews.
However, some of the answers (e.g. Question 60: Classify the predators according to the level of
problem) may have been exaggerated as respondents were made aware that my research focus
was primarily on cheetahs. Furthermore, respondents may have supplied answers when they did
not actually understand the question or had little knowledge on the topic, such as the occurrence
of wildlife abundances. The way the questionnaire was structured, with the use of ground-truthing
questions, reduces this bias (White et al. 2005) but nonetheless care must be taken when
reviewing these answers.
5.2.4 Data analysis
Descriptive statistics (means, standard deviations, frequencies and ranges) were used to explain
the various results. Furthermore, the responses to certain questions/statements were used to
calculate index scores by allocating values of +1, 0 and -1 to the different statements depending
on the responses (Foddy 1994; Walpole & Goodwin 2001). An attitude score was calculated for
each respondent and this represented their overall attitude towards predators (Walpole & 81
Goodwin 2001; Parker et al. 2014; Thorn et al. 2014). The attitude index was based on the sum
of the scores of six relevant questions/statements. Positive responses received a +1, negative
responses received -1 and ambiguous or uncertain answers received a 0 (Walpole & Goodwin
2001). For example, respondents were asked “Do you think wildlife is a natural resource to be
protected?” If they answered yes, they received +1, if they answered no, they received -1 and if
they were uncertain, they received zero. Thus, respondents with higher scores generally had a
more positive overall attitude towards predators. A possible maximum value of +6 and a
minimum value of -6 could be attained.
The same approach was adopted to generate a husbandry and a knowledge index. The husbandry
index was calculated based on the answers given to 29 questions/statements that were designed
to assess the suitability of livestock husbandry methods which were employed by respondents to
protect and manage their livestock. Non-lethal methods of protecting livestock accrued a positive
score (i.e. using herders or guard dogs, fetching livestock from the field, burning fires around the
kraal), whereas lethal methods (i.e. poisoning carcasses or hunting predators) would accrue a
negative score. The index scoring system also took into account the actions taken by respondents
when livestock was lost to a predator, the use of a calving season and maternity kraals, accurate
record keeping, and the kraaling of livestock at night. In the instances where farmers did not have
cattle or small stock (goats and sheep) the response was left blank and not included in the
calculations. The husbandry index was the sum of points obtained for each question/statement
and could accrue a maximum value of +29 and a minimum value of -29. A high index score
would indicate a better approach at protecting livestock and good farm management practices.
The knowledge index was calculated based on the amount of local knowledge of wild predators,
including the current legislation regarding the protection of cheetahs and other wildlife. A high
index score reflected a better understanding of the role and importance of predators in
ecosystems. Furthermore, two posters depicting eight common large predators were used and
respondents were asked to correctly identify each predator (Appendix VII). A total of 14
questions/statements were used to generate the knowledge index score, with a possible maximum
value of +14 and a minimum value of -14.
5.2.5 AIC analyses - attitude index
To assess the relative contribution of different predictor variables on the three indices, I employed
a model building approach using Akaike’s Information Criterion (AIC; Akaike 1974; Burnham
& Anderson 2002). The second-order Akaike’s Information Criterion (AICc) was used for the
dataset to accommodate for the small sample size (Burnham & Anderson 2002). Eight predictor 82
variables were used to assess their potential effect on respondent attitude. The predictor variables
tested included three categorical variables; education level (four levels), primary source of
income (three levels), and position held on the cattlepost (three levels), and five continuous
variables; age, the distance of the cattlepost to the NOTUGRE fence line (km), the total number
of livestock lost, knowledge and husbandry indices. Two respondents were removed from the
dataset prior to the analysis because three or more questions had not been answered. A total of
78 valid respondents were thus used in the overall multi-model analysis to identify the relative
importance of the eight predictor variables on attitude.
Husbandry index
Nine predictor variables were used to assess their potential impact on the husbandry index. I
included the following socio-demographic predictor variables: age, the number of years
respondents had lived in the area, the level of education (categorical variable; four levels) and
the primary form of income (categorical variable; three levels). The total number of livestock lost
and the distance (km) of the cattlepost from the reserve fence-line, characteristics of kraal design
- materials used (categorical variable; five levels) and maximum gap size (cm), and common
circumstances of depredation (categorical: two levels - inside or outside the kraal) were also
included as predictor variables. Fifteen respondents were removed from the dataset as one or
more values were missing, leaving a total of 65 respondents for the analysis.
Knowledge index
Six predictor variables were used to predict the knowledge index of respondents. Five continuous
variables (husbandry index, distance from the reserve fence-line (km), age, number of years
respondents had lived in the area, and total number of livestock lost to predators) and one
categorical variable, education (four levels), were used. Three respondents were removed from
the dataset due to missing values leaving a total of 77 respondents for the analysis. Prior to the
model building, Shapiro-Wilk Normality Tests were conducted to test for normality, and
generalised Variance Inflation Factors (VIF) were used to detect possible multi-colinearity for
all predictor variables (Freckleton 2011). Variables that had a VIF of > 5 were removed to resolve
any co-linearity (Freckleton 2011). I conducted Generalized Linear Models (GLM) to test for the
best combination of predictor variables for the indices. Delta AICc (ΔAICc) values and Akaike
weights (wi) were calculated for each model and were used to explain the strength of each model
relative to the other models and to assess the importance of the individual predictor variables
(Burnham & Anderson 2002). The best model is expressed as the model with the lowest ΔAICc
value. However, any model with a low (< 2) ΔAICc value indicates that it may be suitable
(Burnham & Anderson 2002). Akaike weights also provide a measure of strength of evidence, 83
with higher values indicating the better model suitability (Rowe 2009). Thus, all models with
ΔAICc values < 2 were used, the predictor variables that featured in these models were then
selected and used in a cross validation GLM to identify the best predictor variable/combination
of predictor variables. All analyses were conducted using the statistical software program R
version 3.0.2 and RStudio version 0.98.501 with the packages “car” and “MuMIn” (R Core Team
2013).
5.3 Results
5.3.1 Demographics of respondents:
A total of 80 respondents on cattleposts were interviewed which, to the best of my knowledge,
represented all of the cattleposts within a 14 km buffer of NOTUGRE. Cattleposts that were abandoned
or that did not own livestock were not included or interviewed. One of the questionnaires had to be
removed as the respondent was inconsistent in his answers, thus 79 valid questionnaires were
used for the analyses. All respondents were Motswana nationals and of black African descent.
Seventy-two percent of respondents were female and 27.6% male. On average, each cattlepost
housed 4.6 ± 3.5 persons (range: 1-15). More than 40% of respondents were in the age group of
over 50 years. Less than 8% were younger than 21, with the average age being 47 ±16.5 years
old (range: 14-77 years old) (Figure 5.4). A third of the respondents (35.4%), had no form of
education, and there was a steady decrease in the number of respondents who had any higher
levels of education. Only 5.1 % of respondents had some form of tertiary education (Figure 5.5).
Figure 5.4 The proportion of respondents (n = 79) in each of the five age group categories.
84
Figure 5.5 The highest education level attained by respondents (n = 79) living on cattleposts in communal farmlands west of NOTUGRE, expressed as percentages.
Livestock pastoralism was the primary source of income for the majority (62.7%) of respondents
(Figure 5.6). Some respondents were employed by NOTUGRE and small proportions relied on
other sources of income such as basket weaving, crop farming, palm beer brewing and
government pensions (Figure 5.6).
Figure 5.6 Main sources of family income for respondents living on cattleposts along the western boundary of and within NOTUGRE.
85
5.3.2 Livestock husbandry
A total of 6987 individual heads of livestock were owned by respondent’s households. Each
cattlepost had, on average, 53.13 ± 66.64 goats (range: 5-361), 21.26 ± 42.14 cattle (range: 0-
280), 10.84 ± 16.95 sheep (range: 0-90), 3.1 ± 3.63 donkeys (range: 0-17), and 0.08 ± 0.58 horses
(range: 0-5). The mean number of domestic dogs owned by each respondent was 2.99 ± 2.55
(range: 0-16), and chickens (10.74 ± 10.05/respondent; range: 0-50) also formed an important
part of the agricultural livelihoods on the cattleposts. Grazing land outside of the reserve is
typically communal and mostly unfenced. Livestock were typically not fetched by a herder from
grazing in the afternoon but left to return on their own. Only 3.8% of farmers had cattle in fenced
fields. During the day cattle were almost always left unattended (92.5%) with only 3.8% of
respondents using herders to accompany the cattle (Table 5.1). Kraaling of cattle at night was
used by 64.2% of respondents, with about a third (35.9%) leaving their cattle out of the kraal and
unprotected.
Only one respondent kept his herd of small stock in a fenced field (Table 5.1). Eighteen percent
of respondents used guard dogs to protect their small stock during the day, while the majority
(81.3%) of farmers left their small stock unattended. Kraaling of small stock at night was used
by all respondents and 20.0% of respondents placed livestock guard dogs (LGD) together with
the small stock in the kraals. Most LGDs were local Tswana mixed breeds of medium size (11-
25kg) and all respondents that used dogs perceived them to be effective at protecting the livestock
from depredation.
Table 5.1 Summary of livestock management practices for cattle and small stock during the day
and at night in communal farmlands west of and within NOTUGRE shown as a percentage of the
number of respondents who owned livestock (cattle n = 53; small stock n = 79).
Day Cattle Small stock Fenced field/kraal 3.8% 1.3%
Herder and guard dog 0.0% 1.3% Herder 3.8% 0.0%
Guard dog 0.0% 16.3% Free roaming (unattended) 92.5% 81.3%
Night Kraal and guard dog 1.9% 20.0%
Kraal 62.3% 80.0% Free roaming (unattended) 35.9% 0.0%
86
5.3.3 Kraal design
Cattle kraals consisted mostly of horizontal poles (split-rail fence) with large gaps (38.1%) or
were fenced enclosures (33.3%; Figure 5.7). Other kraal designs included vertical wooden posts
(14.3%), acacia branches (7.1%), or a combination of posts and fencing (7.1%; Figure 5.7).
Illustrations of the different kraal designs are shown in Appendix VIII. The average distance of
cattle kraals from the homestead was 54.58m ± 54.83 (range: 0-1000m), the average height of
kraals was 1.52 m ± 0.43 (range: 0.3-2.5m) and the maximum gap size between individual poles
was 38.95cm ± 31.04 (range: 0-100cm).
Figure 5.7 The different cattle kraal designs and materials utilised by livestock farmers for protecting their cattle at night. Percentages were calculated based on the total number of respondents who owned cattle (n = 53).
Small stock kraals were mostly (52.7%) made using vertical wooden posts (Figure 5.8). Fencing
(13.5%), fencing with diamond mesh (9.5%), acacia branches (4.1%), horizontal poles (4.1%) or
a combination of any of these (16.2%) (Figure 5.8; Appendix VIII). The average distance
between small stock kraals and homesteads was 80.75m ± 186.46 (range 0- 1000m). Kraals were
built at an average height of 1.43m ± 0.46 (range: 0.5-2.5m) with maximum gaps of 11.92cm ±
9.15 (range: 0-35cm).
87
Figure 5.8 The proportions of six categories of kraal designs and the material utilized by livestock farmers (n = 79) for protecting small livestock at night.
5.3.4 Wildlife and predators
Respondents were asked to give their opinion on the status and trends of wildlife (specifically
predator) species that exist in the area. Baboons (Papio ursinus) and honey badgers (Mellivora
capensis) were included as predators as they were frequently mentioned as important predators
of livestock and poultry. Overall, knowledge of the occurrences of wildlife species was poor with
most respondents unsure of their occurrence and/or general trends over the last ten years. Eland
(Tragelaphus oryx), zebra (Equus burchellii), and wildebeest (Connochaetes taurinus) were
considered absent by more than 80% of respondents. Steenbok (Raphicerus campestris), scrub
hare (Lepus saxatilis) and Helmeted Guineafowl (Numida meleagris) were described as being
common by > 43% of respondents. However, very few respondents had an opinion on the trends
in wildlife populations (64.9%) and of those that gave an answer, most indicated that wildlife
populations were stable (13.5%) or increasing (12.8%), although increasing populations were
often based on the presence of juvenile animals and not on the average number seen over time.
These figures included respondents that had cattleposts within the game reserve boundary (n =
11) and where wildlife is likely to be more abundant. In order to have a better understanding of
wildlife occurrence outside of the reserve, I excluded the responses for cattleposts that occurred
within the reserve. Table 5.2 provides a summary of the responses for wildlife occurrence and
trends excluding cattleposts that were within the reserve boundary.
88
Table 5.2 Summary of reported local abundance of wildlife species and population trends over
the last 10 years (n = 69). Wildlife details and trends exclude cattleposts within the game reserve
fence (n = 11).
Status Kudu Impala Eland Zebra Wilde- beest
Duiker Steenbok Warthog Hare Guinea- fowl
Absent 46.8% 25.3% 79.7% 78.5% 81.0% 10.1% 17.7% 49.4% 2.5% 11.4% Rare 20.3% 16.5% 3.8% 2.5% 2.5% 16.5% 15.2% 17.7% 3.8% 20.3% Common 12.7% 32.9% 0.0% 2.5% 0.0% 40.5% 39.2% 15.2% 36.7% 38.0% Very common
3.8% 8.9% 0.0% 0.0% 0.0% 16.5% 11.4% 1.3% 40.5% 13.9%
Don't know
2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5%
Trends Increasing 11.6% 17.4% 2.9% 4.3% 2.9% 17.4% 15.9% 8.7% 24.6% 21.7% Decreasing 14.5% 17.4% 1.4% 5.8% 7.2% 8.7% 8.7% 7.2% 10.1% 7.2% Stable 8.7% 15.9% 14.5% 14.5% 15.9% 18.8% 11.6% 11.6% 10.1% 13.0% Don't know
65.2% 49.3% 81.2% 75.4% 73.9% 55.1% 63.8% 72.5% 55.1% 58.0%
Large predators were believed to be mostly absent on communal farmland apart from spotted
hyenas, baboons and black-backed jackals which were seen on an almost a daily basis (Table
5.3). The majority of respondents (73.0%) were unsure of predator population trends, except for
spotted hyenas and black-backed jackals which were believed to be increasing (Table 5.3). The
presence and trends of predators were mostly from visual observations (49.9%), followed by
tracks (30.6%) and then calls (19.5%).
Table 5.3 Summary of the frequency of sightings of the common predator species seen by
respondents. Perceived predator trends over the last ten years are also included, given as a
percentage of all responses. Answers from all cattleposts were included (n = 79).
Frequency of sightings
PREDATORS Lion Cheetah Leopard Spotted
hyena Brown hyena
Wild dog
Baboon Black- backed Jackal
Honey badger
Never 47.4% 71.8% 48.7% 3.9% 50.7% 92.2% 21.8% 11.7% 25.0% < Once/year 10.3% 9.0% 3.8% 0.0% 4.1% 2.6% 0.0% 0.0% 10.0% Once/year 12.8% 6.4% 3.8% 0.0% 5.5% 0.0% 3.8% 0.0% 8.3% A few times a year
7.7% 3.8% 9.0% 1.3% 1.4% 2.6% 2.6% 0.0% 8.3%
Every few months
11.5% 5.1% 3.8% 1.3% 4.1% 2.6% 1.3% 1.3% 10.0%
Once/month 7.7% 1.3% 14.1% 1.3% 2.7% 0.0% 10.3% 9.1% 10.0% Every week 1.3% 2.6% 6.4% 19.7% 5.5% 0.0% 11.5% 15.6% 18.3% Everyday 1.3% 0.0% 10.3% 72.4% 26.0% 0.0% 48.7% 62.3% 10.0% Trends Decreasing 15.2% 6.3% 8.9% 1.3% 1.3% 6.3% 5.1% 5.1% 1.3%
89
5.3.5 Livestock predation
Respondents were asked to provide information on incidents of predator attacks on livestock over
the preceding 18 months (Table 5.4). Sixty seven out of 79 (84.8%) respondents claimed to have
lost a total of 685 livestock to predators in 704 separate incidents. Furthermore, predators were
also involved in the depredation of chickens and domestic dogs.
Table 5.4 The composition of livestock owned over the survey period and livestock losses to
predators over the preceding 18 months (Jan 2011 – June 2012). A conservative approach was
used whereby if an unknown number of livestock was lost the incident was excluded from the
total count. Injured livestock were included in the counts as they often did not survive.
Livestock Goat Cattle Sheep Donkey Horse Total
Livestock owned 4197 1701 835 248 6 6987 Total livestock lost 556 63 31 52 0 702
(11.7%) (3.6%) (3.6%) (17.3%) (0%) (9. 1%) Losses attributed to predators
Hyena* 154 41 1 6 45 0 238 Leopard 37 7 6 0 0 50
Lion 10 14 0 7 0 31 Cheetah 30 1 0 0 0 31
Wild dog 4 0 0 0 0 4 Baboon 101 0 0 0 0 101
Black-backed jackal 141 0 9 0 0 150 Caracal 18 0 0 0 0 18
Honey badger 2 0 0 0 0 2 Bird of prey 130 0 0 0 0 13 Unknown predator 46 0 0 0 0 46
*the name hyena is used for both species (spotted hyena and brown hyena) as respondents often did not specify although it is likely that they were referring to spotted hyenas which are far more common.
Additionally, a goat was killed by an elephant (but this was not considered as predation), five
domestic dogs were killed by brown hyenas and one domestic dog was killed by a leopard. The
mean number of livestock lost per cattlepost was 8.78 ± 12.57 (range: 0-92) for the 18 month
period, with a monthly average of 0.49 livestock lost per cattlepost. The number of livestock lost
represented approximately 9.1% (range: 0-79%) of total livestock owned (total number of
livestock owned = current number of livestock + livestock depredated). Sixteen respondents
suffered losses exceeding 25% of the total number of livestock owned. The highest percentage
of losses suffered was 79.2% (19 animals).
Hyenas were most frequently blamed for depredation events with a total of 238 livestock believed
to have been killed (33.9%), followed by black-backed jackal in 150 incidents (21.4%), baboons
90
in 101 incidents (14.4%), and 51 leopard incidents (7.1%). Cheetahs were reported to have been
responsible for 31 incidents (4.4%). More specifically, results from incident analysis showed that
hyenas allegedly contributed to 65.1% of all cattle killed, 86.5% of all donkeys and 28.7% of all
small stock depredation incidents. Although hyenas were believed to be responsible for the
majority of attacks on all livestock types, black-backed jackals were also blamed for a number of
attacks of small stock (25.6%) and baboons contributed to 17.2% of small stock attacks. Black-
backed jackals are capable of attacking smaller livestock, particularly lambs and kids, left
unattended and large male baboons were mostly responsible for killing lambs and kids in kraals
that were left unprotected during the day (Respondents, pers. comm.). Other carnivore species
that were involved in incidents of livestock depredation (including poultry) included lions,
caracals (Caracal caracal), birds of prey (Accipitridae spp.), honey badgers, African wild dogs,
civets (Civettictis civetta), mongooses (Herpestidae spp.), and African wild cats (Felis silvestris
lybica). However, predator culpability may have been questionable as the methods used to
identify the predator responsible for the predation event were mostly through the evidence of
tracks (45.5%) or visual sight of the species (27.5%). Killing bites and feeding style, which are
the more acceptable techniques, were rarely used as means of identification (15.7%).
Interestingly, 8.4% of incidents were blamed on predators with no supporting evidence as the
livestock was simply missing.
Perceived predator threat reflects the results of incidents of livestock depredation. Respondents
were asked to rank predators according to most problem causing species. Hyenas were identified
as the most problematic predator by the majority of respondents (54%), black-backed jackals
were classified as the highest ranking by 20.0% of respondents followed by baboons at 15.0%.
Livestock loss was evaluated in terms of economic loss based on the following livestock values.
Cattle were valued at 3500 BWP (Botswana Pula) (367 US Dollar USD) by respondents, goats
at 500 BWP (53 USD) sheep at 700 BWP (73 USD) and donkeys at 400 BWP (42 USD). The
total value of the livestock lost in the preceding 18 months was therefore valued at 540 500 BWP
(56 699 USD). However, reliable record keeping on livestock losses was poor and most (95.5%)
of the records were from memory. Smallstock were most frequently depredated and accounted
for 55.4% of total economic loss. The total cattle loss was valued at 23130 USD representing
40.8%, yet the number of cattle depredated only made up 9.0% of total livestock loss. Table 5.5
tallies the economic value for livestock depredation by each predator species and gives a
representation of the contribution by each predator. In terms of economic value hyenas were
responsible for 46.1% of the total economic losses to all predators, which is estimated at 26 141
USD. Black-backed jackals were found to be responsible for 14.2% of total economic costs and, 91
interestingly, lions were the third most important predator accounting for 10.5 % of the total
economic losses to all predators.
Table 5.5 The total economic costs and contribution (%) of total livestock depredation, by each
predator species and livestock type. Values are presented in US$ at 1US$ = 9.53 BWP (Mid-
market rates: 2014-11-16 08:29 UTC).
Livestock Goat Cattle Sheep Donkey Horse Total %
Hyena 8025 15053 1175 1888 0 26141 46.1 Leopard 1941 2570 441 0 0 4951 8.7
Lion 525 5140 0 294 0 5958 10.5 Cheetah 1574 367 0 0 0 1941 3.4
Wild dog 210 0 0 0 0 210 0.4 Baboon 5297 0 0 0 0 5297 9.3
Black-backed jackal 7395 0 661 0 0 8056 14.2 Caracal 944 0 0 0 0 944 1.7
Honey badger 105 0 0 0 0 105 0.2 Bird of prey 682 0 0 0 0 682 1.2 Unknown predator 2413 0 0 0 0 2413 4.3 29110 23130 2276 2182 0 56698
5.3.6 Circumstances of predation
There was no marked seasonal difference in the number of incidents of livestock depredation
between the cool dry (54. 5%) and hot wet (45.5%) seasons. The month of June had the highest
recorded number of attacks, but care must be taken when interpreting these figures because poor
record keeping means that more recent predation incidents were probably better remembered.
Forty nine percent of depredation incidents occurred at night, of which the majority were outside
the kraal (91.3%) and altogether 88.8% of incidents, both day and night, occurred outside of the
kraal. Yet, all respondents indicated that they kept their small stock in enclosures at night and
62.3% of the farmers who owned cattle indicated that they kept their cattle in enclosures at night.
However, many respondents subsequently indicated that depredation outside the kraal at night
occurred when livestock had not returned to enclosures on these specific nights.
Respondents were asked about common circumstances of depredation events. The majority of
respondents confirmed that most incidents happened at night (58.1%) and outside the kraal
(86.1%); whereas only 28.4% of respondents perceived depredation to be primarily during the
day, and 11.1% respondents found predation to be mostly inside the kraal. Fifty-nine percent of
respondents believed livestock loss to be seasonal of which the cool dry season was regarded to
be the most common season for predator attacks on livestock (60.0%).
92
5.3.7 Trends in perceived predation levels
Respondents were asked to describe the level of conflict with predators over the last ten years.
Many respondents felt that depredation incidents were increasing (38.0%) (Table 5.6).
Respondents were asked to explain why they felt that human-predator conflict was increasing,
most were unsure but some felt it was due to an increase in predator abundances. However, three
respondents mentioned that an increase in depredation events was due to the lack of grazing
resulting in weaker livestock which make for easier prey for predators. To compensate for the
poor grazing, many farmers would leave the livestock out at night to have more grazing time. Of
the respondents who indicated that conflicts were decreasing (24.1%), improved farm
management was reported by five respondents as the reason for changes and two respondents
attributed the decrease in conflict to a decrease in predator abundances.
Table 5.6 Summary of the perceived trends in livestock depredation over the last ten years (n =
79). The first column N is the number of respondents and the second column is the percentage of
all respondents.
Trends N Percentage
Increasing 3 38.0%
Decreasing 1 24.1%
Stable 2 27.8%
Unsure 8 10.1%
5.3.8 Predators removed by farmers
Respondents were asked if they had removed (killed or caught) predators in the past ten years.
Only seven of respondents admitted to having removed predators. A cheetah had been killed by
hunting with domestic dogs, a number of hyenas (numbers unspecified) had been caught in gin
(leg-hold) traps and subsequently killed, a number of black-backed jackals and hyenas (numbers
unspecified) were killed by hunting with dogs, and a leopard had been shot with permission of
the Department of Wildlife and National Parks (DWNP).
5.3.9 Livestock loss
Overall, respondents classified drought (67.5%) as their biggest problem when it came to their
livestock farming, followed by predators (60.0%) and diseases (53.8%). However, insufficient
grazing was also expressed as a major problem by a number of respondents (47.5%). Other
problems encountered by farmers were infertility, poor quality grazing, low yields, unreliable
93
market, theft, snares, veterinary cordon fences and miscarriages. Insufficient grazing was
identified as the second most important concern by about two-thirds (66.7%) of respondents who
did not express it as the biggest problem.
5.3.10 Knowledge of local predators
Less than half the respondents could correctly identify a cheetah (48.8%). Leopards were
correctly identified by 57.5%. Leopards and cheetahs were often confused with each other. By
contrast, 77.5% of the respondents could identify lions and 75.0% a black-backed jackal. Wild
dogs and brown hyenas could only be identified by 43.8% and 31.3% of respondents,
respectively. Overall, correct identification of all eight predators was very poor, with less than
half of the respondents capable of identifying all of the common predators. Interestingly,
respondents who had scored highly on predator identification were asked where they had learnt
to identify these predators. It was found that those with family members who work, or used to
work, in game reserves or the like (e.g. captive facilities) had a particularly high score. Fifty four
percent of respondents that correctly identified all predators knew someone that worked in a
reserve.
5.3.11 Attitudes and perception towards wildlife
Sixty nine percent of respondents believed that cheetahs should be protected in Botswana but
despite this, only 38.5% of respondents attached any positive value to the cheetah. Positive values
included ecotourism, beauty, employment, and the need for future generations to see cheetahs.
Only 20.0% of respondents had a positive attitude towards sharing the land with predators, 59
(73.8%) had a negative response and five respondents (6.3%) were either indifferent or unsure.
Despite the negative perception of the coexistence of predators on farmland, the majority of
respondents (80.0%) agreed that wildlife should be protected as it is a national resource. In
response to the question, “who do you think is responsible for the predator-livestock conflict?”
53 out of the 79 respondents (67.1%) believed it to be the responsibility of the Botswana
government. Seven respondents felt it was the responsibility of the owner of the livestock, whilst
six respondents held the game reserves responsible. Other parties that were mentioned as
accountable were non-governmental organizations (NGO) and everyone, whilst four respondents
were unsure. There were mixed opinions regarding the solution to the protection of wildlife.
Respondents primarily mentioned translocation as a solution towards the coexistence of predators
on farmland (35.0%). Improved farm management was expressed by 20.0% of total responses.
94
5.3.12 Attitude, knowledge and husbandry Indices
Predictors of Attitude
The mean attitude index score was -0.29 ± 2.27 (range: -4 to 6). A large proportion of respondents
(46.3%) had an attitude index of -1 or less and only 8.8% of respondents had an attitude index
with a positive value of three or more, indicating an overall negative attitude towards predators
on farmlands. The attitude index had a low Cronbach reliability α = 0.35 (range: 0.26-0.37). The
Cronbach’s alpha value is low in relation to the acceptable reliability value of 0.70 (Santos 1999).
However, lower thresholds are sometimes used and reported in the literature (Santos 1999). All
questions/statements produced similar values so no single question/statement could be removed
to improve (increase) the overall reliability of the index. The low alpha value may be the product
of a limited number of questions/statements used (n = 6) to construct the index (Gliem & Gliem
2003).
A GLM identified 17 out of 256 potential models to best explain the attitude index (ΔAICc value
< 2). These models included six predictor variables; knowledge index, husbandry index, distance
to the reserve, respondent’s position, highest level of education, and primary form of income
(Table 5.7). However, the cross-validation GLM (i.e. a GLM that only included the six variables
identified above) indicated that no one predictor variable/combination significantly influenced
the attitude indices of the respondents.
Table 5.7 The top 17 models and identified variables used to predict the attitude index. Models
are arranged in descending orders according to their AICc scores. ‘Education’ refers to the
highest level of education, ‘Income’ is the main source of family income, “Position” explains the
position (owner/employee) held on the cattlepost, ‘Distance’ describes the kilometres to the
boundary of the reserve, ‘Husbandry’ refers to the Husbandry Index score, and ‘Knowledge’ is
the Knowledge Index score.
Model Education Income Position Distance Husbandry Knowledge AICc ΔAICc AICc weight
77 X X X 337.4 0.00 0.038 109 X X X X 337.5 0.05 0.037 101 X X X 337.5 0.10 0.036
69 X X 337.7 0.33 0.032 89 X X X 338.1 0.65 0.028 73 X X 338.2 0.81 0.025 65 X 338.2 0.81 0.025 67 X X 338.5 1.04 0.023
95
93 X X X X 338.5 1.12 0.022 75 X X X 338.7 1.27 0.020 125 X X X X X 338.7 1.27 0.020 81 X X 338.7 1.32 0.020 117 X X X X 339.0 1.54 0.018 91 X X X X 339.0 1.63 0.017 83 X X X 339.1 1.73 0.016 85 X X X 339.2 1.74 0.016 71 X X X 339.3 1.91 0.015
Predictors of Husbandry
The husbandry index scores were generally positive, with a mean score of 3.3 ±3.5 (range -4 to
14). The index had an acceptable internal consistency (Cronbach reliability α = 0.56; range =
0.63-0.71). Nine predictor variables were used to test their effects on the husbandry index.
Six models (ΔAICc value < 2) best described the data and these included four of the predictor
variables; perceived common circumstances of attack, main form of family income, maximum
gap size recorded in the kraal walling, and total number of livestock loss to predators in the
previous 18 months (Table 5.8). The ‘best’ model (model 67) included both circumstances of
attack and gap size (wi = 0.065). However, the cross validation GLM indicated that no one
predictor variable/combination significantly influenced the husbandry indices of the respondents.
Table 5.8 Summary of the top six models and four predictor variables that best described the
dataset for the husbandry index. The models are arranged in descending order based on their
AICc values. ‘Circumstances of attack’ refers to the common circumstances of depredation
incidents on livestock, ‘Income’ is the respondent’s main source of income, ‘Gap size’ is the
largest size gap (cm) recorded in the kraal walling, and ‘Livestock loss’ refers to the total number
of livestock loss to predators in the previous 18 months.
Model Circumstance of attack
Income Gap size Livestock loss
AICc ΔAICc AICc weight
67 X X 338.1 0.00 0.065 3 X 338.2 0.09 0.062
73 X X 338.8 0.69 0.046 65 X 339.4 1.28 0.034
201 X X X 339.5 1.43 0.032 195 X X X 339.9 1.86 0.026
96
Predictors of Knowledge
The mean knowledge index was 2.76 ± 5.25 (range: -9 to 14) (Cronbach reliability α = 0.69;
range = 0.45-0.59). Five models best described the data, with four predictor variables (age of
respondents, the highest level of education, the distance of the cattlepost to the reserve boundary,
and the number of years respondents had been living in the area) (Table 5.9).
Model 22 had the highest Akaike weight (wi = 0.118) and included age, distance and livestock
loss (Table 5.9). However, the cross validation GLM indicated that no one predictor
variable/combination significantly influenced the knowledge indices of the respondents.
Table 5.9 Summary of the five top models and variables that best predicted the knowledge index.
The models are in descending AICc order. ‘Age’ refers to the respondent’s age in years,
‘Education’ is the highest level of education obtained, ‘Distance’ is the cattlepost’s distance (km)
from the boundary of the reserve, and ‘Years’ refers to the number of years a respondent has
lived in the area.
Model Age Education Distance Livestock loss Years AICc ΔAICc AICc weight
22 X X X 470.1 0.00 0.118 21 X X 470.2 0.09 0.112 6 X X 471.5 1.46 0.057
54 X X X X 471.9 1.81 0.048 23 X X X 471.9 1.86 0.046
5.4 Discussion
Questionnaire interviews can be used to evaluate interactions between humans and predators,
providing measurable data that can be quickly and relatively cheaply collected (Holmern &
Røskaft 2013). However, caution needs to be exercised when reviewing results as questionnaires
rely on information that is subjective and sometimes misleading. For example, the exaggeration
of livestock losses and bias towards certain predator species (Graham et al. 2005; Holmern &
Røskaft 2013). Furthermore, the relatively small dataset used in my study, despite including all
available cattleposts, also likely affected the results making it difficult to determine relationships
between ecological and social variables.
5.4.1 Livestock husbandry practices
Livestock protective methods in the communal farmland consisted mostly of the kraaling of
livestock at night. However, most livestock herds were not fetched in the afternoon and left to
97
return on their own. Furthermore, some farmers did not count their livestock upon return so any
strays would sleep out at night. During the day livestock herds were mostly left unattended and
unprotected with relatively few farmers employing herders and/or LGDs to protect their
livestock. This is despite respondents considering LGDs to be effective. The use of LGDs has
been documented as an effective tool to reduce predation (Coppinger et al. 1988; Marker et al.
2005; Gonzalez et al. 2012). Only 16 respondents used LGDs and these were only used to protect
small stock, yet a total of 239 domestic dogs (mean and SD: 2.99 ± 2.55 per cattlepost) were
owned on cattleposts. The relatively low number of farmers that used LGDs as a protective
method suggests that farmers are either not aware of the benefits of using dogs to protect
livestock, or do not know how to train a dog to become a livestock guardian (pers. obs.).
Livestock farmers in the communal farmlands may therefore benefit from a formal workshop on
the use of LGDs including instructions on livestock guard dog training. Furthermore, subsidising
of dog food may be an incentive for farmers to adopt this protective method (pers. obs.).
Four main livestock kraal designs were identified and consisted of tightly fitted mopanewood
(Colophospermum mopane) posts, split-rail fencing, tightly packed acacia branches, or wire
fencing. Some farmers also had maternity kraals where young goats and sheep would be kept
separately during the day. All kraals that were built with the use of acacia branches had the stems
of the branches facing out of the kraal. Additionally, the walling of all kraal designs, particularly
for cattle, typically had large gap sizes suggesting that kraals are built to prevent livestock from
escaping rather than to prevent predators from entering the kraals (see also Patterson et al. 2004).
Nonetheless, kraaling livestock at night, regardless of the kraal design, appears to be effective in
reducing livestock losses as livestock depredation was mostly recorded when farmers failed to
kraal their livestock at night. Similar findings have been recorded in previous studies (Schiess-
Meier et al. 2007; Kgathi et al. 2012; Parker et al. 2014). Large predators have been found to
avoid close proximity to human settlements (Ramakrishnan et al. 1999; Ogada et al. 2003;
Pettorelli et al. 2009). Therefore, keeping livestock at night close to the homestead, regardless of
the kraal design, may decrease depredation incidents. Some farmers took added precaution by
building fires around the kraals or building some form of roofing on maternity kraals to prevent
baboons from entering the enclosures during the day when homesteads were unoccupied. Other
protection methods utilised included raised chicken pens to protect chickens from honey badgers.
These findings suggest that, where protective measures are used, they are effective but many
farmers lack a proactive approach towards the raising of their livestock. Regaining farmers’ self
responsibility for their livestock may change overall husbandry effort, as active defence and herd
attendance are essential measures of animal husbandry (Patterson et al. 2004).
98
Despite the higher value of cattle, husbandry methods were typically less intensive for cattle than
for smaller stock. Cattle were mostly left to roam freely and unattended. Furthermore, one farmer
indicated that he had not seen his cattle in over two weeks as they had wondered off in search of
better grazing. However, despite a lower husbandry effort, cattle did have a lower depredation
rate than goats (cattle = 3.7% goats = 13.0%). Some farmers had expressed that because of a lack
of grazing they had recently moved their cattle to other areas, whereas other farmers were
allowing cattle to graze for longer hours sometimes leaving them to graze out during the night.
Thus, the relaxed protective methods observed may have been temporary in reaction to the
drought conditions, and more stringent kraaling may ordinarily be used. This might explain the
lower depredation of cattle. Another possible explanation is that cattle are typically preyed upon
by larger predators such as lions (Patterson et al. 2004; Holmern et al. 2007; Selebatso et al.
2008; Sifuna 2010; Kgathi et al. 2012). Thus, the relatively low depredation of cattle may indicate
that larger predators (i.e. lions) occur at lower densities on farmlands than smaller predators such
as black-backed jackals (Patterson et al. 2004). Indeed, lions were mostly described as absent by
respondents (47.4%) on communal farmlands. Furthermore, certain livestock types may be more
vulnerable than others due to differences in behaviour including herd composition and vigilance
(Polisar et al. 2003). For instance, some livestock may have less flight capability and weaker
defences (Polisar et al. 2003).
Accurate record keeping of past incidents of livestock depredation was typically lacking; this was
also noticed in a previous study in the Ghanzi District in south-west Botswana (Selebatso et al.
2008). In addition, some farmers did not know the exact number of livestock and poultry they
owned. Poor record keeping and irregular livestock inspection may cause farmers to unjustly
blame predators for livestock losses (pers. obs.). Additionally, it appeared that depredation
incidents by larger predators including lions, leopards and cheetahs were more easily
remembered than depredation events by baboons, black-backed jackals and hyenas. Respondents
would often name incidents by larger predators and only report other incidents when we
questioned them on any losses due to other predators. Possible reasons for this are that larger
predators may hunt cattle which have a higher value to farmers, hence incidents are more
noteworthy, and furthermore, livestock losses by these predators accrue compensation.
5.4.2 Wildlife and predators
Accurate information on wildlife (prey availability) and predator abundances is essential to
explain predator-prey interactions (Landa et al. 1999; Graham et al. 2005), as human predator
conflict is often indirectly fuelled by the depletion of wild prey from poaching and competition
with livestock (Graham et al. 2005; Winterbach et al. 2014). There is a lack of data on the 99
abundances of wildlife and predators occurring in the communal farmlands adjacent to
NOTUGRE, but during my extensive fieldwork in these areas I rarely saw wildlife. Respondents
were asked to provide estimations on wildlife occurrences and frequencies of sightings. Larger
species, including eland, zebra, wildebeest and kudu were generally considered to be absent. Only
smaller prey species including steenbok, scrub hare and helmeted guinea fowl were described as
being common. However, these are perceived abundances which are based on a very subjective
evaluation. Significantly, livestock density was calculated at approximately 20.7 livestock/km².
An overexploitation of wildlife (through poaching), coupled with high livestock density and a
corresponding increase in competition for natural resources (food and water), can reduce the
density of wildlife, particularly of large prey (> 60 kg), outside of protected areas (Mishra et al.
2003; Graham et al. 2005; Winterbach et al. 2014). Predators will prey upon wild prey species
in preference to domestic livestock, but where the prey base is absent or limited, predators may
resort to killing domestic livestock (Landa et al. 1999; Patterson et al. 2004; Graham et al. 2005;
Schiess- Meier et al. 2007; Winterbach et al. 2014). Livestock losses are not related to predator
density, but are rather a function of livestock availability (Landa et al. 1999; Graham et al. 2005).
Therefore, reducing predator abundances, in anything less than a radical eradication of isolated
populations, is unlikely to resolve conflict (Landa et al. 1999; Graham et al. 2005). Predators are
more likely to prey on livestock where livestock occurs at higher densities than wild prey (Landa
et al. 1999; Polisar et al. 2003; Graham et al. 2005; Winterbach et al. 2014). However, where
livestock is well looked after, including kraaling during the night and guarding during the day,
and the natural prey base is not depleted, large predators will prey upon wild prey even when
livestock is more abundant (Marker et al. 2003d; Ogara et al. 2010; Winterbach et al. 2014).
Conserving natural prey should not be overlooked when attempting to reduce livestock
depredation in the context of large predator conservation (Polisar et al. 2003; Mishra et al. 2003;
Clarke 2012). The complex ecological interactions require a multi-species and ecosystem
management, thus it is important to also consider the quality of the habitat as this might affect
prey availability (Graham et al. 2005). The severe overgrazing observed in the communal
farmlands is likely to negatively affect the abundance of wild prey species.
5.4.3 Level of conflict
Livestock losses were reported to be caused most often by hyenas, followed by black-backed
jackals and baboons. This was reflected by the respondent’s ranking of predator importance.
Hyenas were also reported to be the most substantial damage-causing predator in the Ngamiland
District in northern Botswana (Kgathi et al. 2012), and near the Serengeti National Park in
Tanzania (Holmern et al. 2007). In a study conducted in Zimbabwe, baboons were reported to
contribute to the majority of goat and sheep predation (Butler 2000). Leopards were not important 100
predators in my study, contributing to 7.3% of all livestock losses and 8.7% of total economic
value. In contrast, leopards were ranked as themost frequent predator in studies conducted in the
Ghanzi District in south-west Botswana (Selebatso et al. 2008) and the Okavango Delta region
in the Ngamiland District in northern Botswana (Sifuna 2010). In terms of economic loss, lions
were found to be the third most important predator, particularly for cattle. A number of studies
also found lions to be important predators of cattle (Patterson et al. 2004; Holmern et al. 2007;
Selebatso et al. 2008; Sifuna 2010; Kgathi et al. 2012). However, these findings may have been
associated with species that accrue compensation, particularly for studies that obtained data from
DWNP Problem Animal reports (Selebatso et al. 2008). Cheetahs attacked mostly smaller prey
(goats) with only one recorded attack on a calf, with a total contribution of only 4.5% of all
livestock losses to predators, which represented 3.4% of the total economic value of livestock
depredation.
The observed high depredation by hyenas and black-backed jackals may suggest a higher
abundance of these species (Patterson et al. 2004; Yirga et al. 2014). However, both species have
distinctive calls and so are more easily detected and identified than other predators (Skinner &
Chimimba 2005). Nonetheless, they are both opportunistic feeders and have behavioural
plasticity and so may prey upon vulnerable livestock and scavenge from livestock carcasses
(Hall-Martin & Botha 1980; Yirga et al. 2014). Due to their ecological flexibility and behavioural
plasticity both of these species are more likely to adapt to anthropogenic landscapes and therefore
be important predators of livestock (Holmern et al. 2007).
The depredation impact (percentage lost) on the livestock in my study constitutes a significant
proportion of the total livestock owned (~9.8%) in comparison to other studies (Graham et al.
2005). In a comprehensive global study on human-predator conflicts, Graham et al. (2005) found
livestock loss to range between 0.02 – 2.6% of all livestock owned. Sixteen respondents in my
study reported livestock depredation that exceeded 25% of their total number of livestock. This
economic impact can be substantial for poor rural subsistence farmers that may only own a few
livestock (Mishra et al. 2003). For instance, loss of one sheep or goat may represent a loss of one
month’s pension for a cattlepost resident. This high economic impact may reduce tolerance
towards predators and provoke retaliatory persecutions (Sifuna 2010; Lindsey et al. 2013;
Chattha et al. 2013). This is particularly so where livestock provides the only means of livelihood
(Dickman 2010).
Tolerance of livestock depredation differed with the different livestock types. Tolerance appears
to be associated to the social value of the livestock, thus the loss of a cow was considered to be 101
substantial however very little social and/or monetary value was attached to donkeys and their
losses were often tolerated and left unreported. Donkeys were the most depredated livestock type
relative to the number owned (21%), whereas cattle were the least (3.7%), yet because of the
social value attached with cattle it is likely that farmers have a lower tolerance towards predators
attacking cattle. Thus, addressing the problem of depredation of more valuable livestock should
be a priority when implementing livestock mitigation measures.
Interestingly, the loss of chickens was often considered to be important and respondents often
commented on the problem of honey badgers and other small predators. Most studies have
focused on the human-large predator conflict yet small- (average body weight <7kg) and
medium-sized predators (7-25kg) may prey upon poultry and even young kids and lambs
(Holmern & Røskaft 2013). Honey badgers, African wild cats, and birds of prey were often
reported as important predators by farmers. Furthermore, baboons and black-backed jackals were
also reported as important culprits in poultry loss. Poultry can have an important nutritional and
financial value for rural farmers (Holmern & Røskaft 2013).
Livestock diseases and poor nutrition due to drought were most frequently identified as the
biggest problems faced by livestock farmers. All interviews were conducted over the dry season
and over a drought, hence the results may have been influenced by prevailing conditions.
Unfortunately, due to poor record keeping, I could not quantify the value of these losses.
However, Graham et al. (2005) found that many studies evaluate livestock losses to other causes
than predators to be proportionally more financially damaging (Graham et al. 2005). Predation
may also mask underlying causes such as poor husbandry including poor diet and health (Graham
et al. 2005). Nonetheless, other causes of livestock loss are often not considered by farmers and/or
are more tolerated.
The presence of a predator does not prove livestock depredation. Predators are sometimes blamed
for missing or stolen livestock (Graham et al. 2005). This was also the case in my study where a
number of respondents held predators responsible for missing livestock. Poor husbandry
practices predispose such behaviour as carcasses are seldom found when livestock is left
unattended and thus predators are blamed with little or no proof. It is therefore difficult to
conclude what percentage of livestock losses is positively a result of depredation without
intensive monitoring.
102
5.4.5 Trends
Many respondents felt that livestock losses to predators have been increasing in the last ten
years. The reason was often said to be the growing numbers of predators although some
farmers expressed that poor grazing quality and starving livestock may have also lead to
predators preying upon the weaker animals. However, increased predation incidents may be
as a result of a depleted prey base and increasing livestock densities. It is also possible that
respondents expressed that the conflict is deteriorating to emphasise the severity of the
problem (Marker et al. 2003c). “Hyper-awareness” is also common; this is where respondents
exaggerate their losses intentionally or unintentionally even where they may not have
personally experienced wildlife conflicts (Dickman 2010). Perceptions of damage causing
predators may come about from only one incident experienced by a community member
(Dickman 2010). For instance, during my interviews a respondent was particularly unhappy
with wild dogs despite never loosing livestock to wild dogs.
5.4.6 Farm management recommendations
The level of human-predator conflict on communal farmlands appears to be high; livestock
losses are extensive and persecution of large predators’ both outside and within the reserve
may have severe consequences on predator populations, particularly on the relatively small
cheetah population (see Chapters 3 and 4). Present livestock husbandry measures appear to be
insufficient for acceptable and tolerable levels of livestock losses. Improving current farm
management and animal husbandry practices, including implementing a proactive attitude
such as daily record keeping, fetching livestock from pastures and ensuring all livestock has
returned and is kraaled at night, will not only reduce incidents of livestock loss due to
predators (Graham et al. 2005), but more importantly, the loss of livestock by theft, snaring,
diseases and starvation should also decrease as farmers will have the opportunity to identify
any sick or injured animal (Schiess-Meier et al. 2007).
Good husbandry practices; livestock accompanied by a herder during the day, kraaling livestock
at night with LGDs reduces livestock depredation and may, in the long term, prevent
predators from becoming habitual livestock hunters (Marker 2002; Ogada et al. 2003).
Kraals need to be predator proofed and built away from dense bushes and in close proximity
to active homesteads (Ogada et al. 2003). Losses can further be reduced by burning fires around
the kraals to deter predators at night (Kgathi et al. 2012), synchronized livestock breeding
seasons and using calving kraals that are well protected (i.e. roofing) and close to human
habitation (Marker 2002; Polisar et al. 2003), and stocking certain breeds of cattle and goats that
are less vulnerable to predation than others (Landa et al. 1999; Polisar et al. 2003). Furthermore, 103
the livestock needs to be in healthy condition and well fed, this may require reducing the
density of livestock on the already overgrazed land and moving the livestock to more arable
land during drought periods. Seasonal management of livestock may further reduce frequency
of predator attacks that are elevated in the dry season. The farming community can also assist
in restoring wild prey populations by ensuring there is sufficient available forage and reducing
poaching.
5.4.7 Management implications
Over an extended period of time (from the 1890s until 1960s) there was widespread eradication
of all large carnivores in the Tuli Block (Lind 1974) as they were seen as vermin by livestock
farmers establishing farms. Consequently, large carnivores have, for the most part, been
extirpated from farmlands within the Tuli Conservation zone (Winterbach et al. 2014).
However, with the establishment of game reserves, large carnivore populations have shown
some recovery in numbers and re-occupation of former ranges has taken place (McKenzie 1990).
Due to the prolonged absence of large carnivores, most traditional husbandry practices have
been abandoned over time (Kgathi et al. 2012). Indeed, livestock guarding in the rural
communal farmlands is limited and farmers lack a proactive approach towards the raising of
their livestock. Conflicts between people and predators are emerging and growing in regions
that are experiencing recovering predator populations after extended periods of local extinction.
But there may be resistance among farmers in readopting some of the traditional husbandry
practices as they are potentially costly (i.e. employing a herder) and require willingness to a
change in lifestyle (Ogada et al. 2003). Children were commonly used as herders in the
past but are now required to attend school (Kgathi et al. 2012). Technical assistance and
economic support, such as subsidy of husbandry practices may encourage farmers to change
their farm management practices and reduce depredation rates, possibly providing the first step
towards mitigating the HWC. Alternatively the DWNP could enforce the use of responsible farm
management (Klein 2007).
5.4.8 Aspects of attitude, knowledge and husbandry
Attitude is considered to be an important aspect of conflict mitigation efforts, with the
prevalent assumption that hostility is directly affected by the level of predation (Dickman
2010). No set of factors best explained the attitude, husbandry or knowledge of respondents
living alongside predators, although the knowledge index was identified as an important
factor in shaping respondents’ attitudes. Previous studies have reported that education and
knowledge are important drivers of attitudes and encourage farmers to be involved in the
104
planning a n d decision-making concerning the management of large predators beyond protected
areas (Bath 1998; Ericsson & Heberlein 2003; Røskaft et al. 2007; Selebatso et al. 2008).
Respondents expressed a negative attitude overall towards the conservation of large predators.
But despite livestock forming an important source of income and food, neither economic
dependency nor the extent of livestock loss influenced the attitudes of respondents. Attitude is
not only shaped by human-wildlife interactions and personal experiences, but it may also be
a product of social factors and human-human conflict (Dickman 2010). Interactions between
people and authorities can play a substantial role in human-predator conflict and is often
overlooked (Dickman 2010). Thus, understanding and improving the relationship between the
local people and conservation bodies such as the DWNP and NOTUGRE is equally important
to effectively mitigate conflict. If residents have had a negative experience, they may view
the reserve or local authorities with a negative attitude which may lead to negative attitudes
towards wildlife conservation. Attitudes towards local authorities (DWNP and Botswana
government) and NOTUGRE were not investigated in this study but some respondents clearly
demonstrated their unhappiness with either the reserve or the local wildlife authorities. Thus, I
feel that improving these relationships is a critical aspect towards shaping more positive attitudes
and should be investigated further.
5.4.9 Compensation implications
Compensation schemes have been implemented in Botswana in an effort to reduce HWC by
increasing tolerance for losses and reducing retaliatory killing of damage-causing wildlife
(Bulte & Rondeau 2005; Jackson et al. 2008; Selebatso et al. 2008; Kgathi et al. 2012).
Furthermore, it has been suggested that reports from compensation schemes can be used to
document the current conflict as farmers are more likely to report livestock losses if they
have financial incentives (Klein 2007; Selebatso et al. 2008). However, critics have argued
that compensation schemes are inefficient in reducing conflict and may even encourage
farmers to relax their protective measures (Bulte & Rondeau 2005; Klein 2007; Clarke 2012;
Kgathi et al. 2012). This is apparent on many cattleposts located outside or within NOTUGRE
where the blame for human-predator conflict has shifted towards the government body and
livestock protection and care was mostly believed to be the responsibility of the government.
Compensation schemes are often inefficient due to a number of challenges associated with
implementation, including a high financial budget and man power required to process the claims
(Jackson et al. 2008; Kgathi et al. 2012). The government is also committed to continuing this
program indefinitely. Farmers tend to only report incidents which accrue financial
compensation, consequently the information gathered from Problem Animal Control (PAC) 105
reports does not necessarily give an accurate picture of the predator conflict as the dataset is
invariably biased in terms of predator species and is likely to under estimate the extent of
livestock depredation (Landa et al. 1999; Schiess-Meier et al. 2007). Assisting farmers to
protect their livestock is believed to be a better solution (Clarke 2012). In 2009, the Botswana
Government updated their compensation policy to include the requirement of adherence to
certain farming management practices (herding of livestock during the day and enclosing
the livestock into well-constructed kraals at night) to avoid potential moral hazards that may
arise from negligent farmers with poor livestock husbandry practices (Bulte & Rondeau 2005;
Kgathi et al. 2012).
In my study, most respondents had suffered livestock losses to predators (83.8%), yet only
about half had previously contacted a wildlife officer. Incidents of livestock depredation were not
always reported to wildlife officers, particularly if it was damage done by a hyena as that would
not warrant compensation under Botswanan legislation. Some respondents expressed their
dissatisfaction with the compensation policy. The current compensation rate for livestock loss
due to predators is approximately 35% of the market value of the livestock (Kgathi et al.
2012). Therefore, it is unlikely that the compensation scheme is effective in terms of alleviating
the human-predator conflict. In addition, the DWNP is responsible for the implementation of
laws against the illegal killing of predators, however these are difficult to enforce due to the
limited man power available and large distances involved (Klein 2007).
The success of compensation schemes relies on a streamlined, adequate, and efficient system.
An incentive program, where the farmers are involved and able to implement decisions
within the community, may gain the support of the community for sustainable coexistence
between farmers and predators (Clarke 2012). Mishra et al. (2003) designed an incentivised
program, a locally managed communal insurance program, where farmers contribute monthly
premiums for their livestock in a communal insurance fund to offset the costs of livestock
losses (Mishra et al. 2003). A similar system could be adopted in the communal farmlands
bordering onto NOTUGRE. The program appoints local community members to supervise
the implementation of the insurance compensation scheme and regulations of the funds are
discussed between the community council and the government body (Mishra et al. 2003).
Initially, the Government and NGOs can help contribute funds into this cooperative fund until it
is self-sustaining (Mishra et al. 2003). Incentives may be provided to encourage good
livestock husbandry by rewarding farmers that have had the least annual number of livestock
losses; producing realistic rates of compensation at 100% the value of the livestock; and
discouraging false compensation claims (Mishra et al. 2003). 106
5.4.10 Outreach
The general consensus among respondents for resolving conflict was to translocate large
predators out of farmlands. Furthermore, respondents often supported the conservation of wildlife
but only within protected areas. A similar predisposition was found in a questionnaire survey in
the Ghanzi District, where significantly fewer farmers supported the conservation of cheetahs
outside protected areas (Selebatso et al. 2008). Promoting the value of wildlife in farmland
ecosystems can increase conservation awareness (Marker et al. 2003a). Equally, understanding
the impact and consequences of persecution, particularly indiscriminate killing by the use of wire
snares and poison, is crucial in the preservation of biodiversity. Environmental education
programs provide a platform to explain the current effect of predator persecution and the
successful non-lethal methods available to reduce the loss of livestock. However, the success of
this program relies on complete transparency from conservation authorities where the purpose of
the program is clearly stipulated from the outset otherwise a negative attitude may be formed
from false, negative and incorrect information given. This is best achieved by providing specific
knowledge such that local communities can make informed decisions. The community may
further benefit from information on local predator species including techniques to identify
culpable predators in an event of predator loss.
Although environmental awareness can improve the overall understanding of the importance of
wildlife, rural farmers may have other priorities (Clarke 2012). Sustainable use of the land for
long term benefit is not necessarily a priority, many farmers live day-to-day and there is little
incentive to protect wildlife which does not give direct financial benefit (Mishra et al. 2003;
Clarke 2012). Economic incentives for the conservation of wildlife on communal farmland may
increase the value of wildlife, such as through the well managed and sustainable consumptive
use of wildlife, and can result in positive attitudes (Klein 2007; Sifuna 2010). It is important
that the local people’s needs and rights are taken into account (Clarke 2012). Economic gains
such as ecotourism and hunting can increase the value of wildlife and hence increase wildlife
tolerance and attitudes towards the coexistence of predators on farmlands (Mishra et al. 2003;
Klein 2007; Sifuna 2010). Increasing wildlife numbers by the banning of hunting has the
reverse effect; Kenya is a prime example of failure, losing 60-70% of all its wildlife since
the ban of hunting and consumptive use of wildlife in 1977 (Clarke 2012). Furthermore,
legalising consumptive and sustainable harvest and giving authorization for communities to
jointly manage their wildlife may reduce poaching which largely comes from communities
that border onto reserves (Sifuna 2010). The survival of wildlife relies upon the support of
107
local communities and consumptive use of wildlife is likely to encourage this support (Sifuna
2010).
Local communities require ownership of natural resources and involvement in decision-
making regarding wildlife management (Bath 1998). However, successful community run
concessions require a flawless operation that is free of corruption and greed so that income
generated from wildlife benefits those affected (Clarke 2012). A successful outcome requires
interest and dedication on the part of the community, but this might be difficult to achieve as
the average farmer does not have the desire to work harder and has few ambitions (Clarke
2012). However, increasing their appreciation for wildlife could gain their support for
conservation initiatives (Hudenko 2012). Positive encounters with wildlife can evoke a positive
emotional response and affection which can positively change the attitude towards wildlife
(Røskaft et al. 2007; Hudenko 2012; Heberlein 2012). Educational programs such as Children
in the Wilderness (CITW) take children from local communities to neighbouring lodges in
protected areas where the children not only learn about the importance of the natural
environment but are also taken on wildlife viewing drives where they have a chance to see
and experience their natural heritage, inspiring them to become custodians of the
environment.
5.5 Conclusion
Livestock losses experienced by farmers in farmlands adjacent to and within NOTUGRE
appear to be relatively high compared to previous studies, but may be a consequence of the
lack of proactive livestock protection measures. Farm management training that includes
preventative measures for livestock depredation, correct techniques to identify the predators
responsible as well as overall improved livestock husbandry would benefit rural subsistence
farmers. Farmers need to gain self-accountability and responsibility for their livestock, which
requires them to better protect their livestock from predators and improve current livestock
husbandry practices. The current compensation scheme was initiated as a measure to mitigate
the human-predator conflict however it does not appear to have resolved the problem and
may even have shifted responsibility. In so doing, the wildlife authorities are perceived to be
accountable for the conflict. Conflict mitigation plans may benefit from a locally managed
communal insurance program that is implemented by the community in collaboration with
the DWNP; improving self-responsibility as well as the relationship between local
communities and wildlife authorities. Improved co-operation may also be achieved by
organising farmers’ meetings to address concerns in the farming community and the DWNP 108
assisting with infrastructural support. HWC mitigation and the coexistence of predators and
necessitates a more positive attitude towards the conservation of predators (Bath 1998).
Human-predator conflict cannot be resolved by reducing the losses of livestock and
understanding the socio-economic environment is crucial to the design and implementation of
successful conservation plans (Dickman 2010). Conflict resolution requires a multi-
disciplinary approach that is specific to the area and includes all stakeholders (Dickman
2010).
109
CHAPTER 6 SYNTHESIS AND CONCLUSION
Prior to my study there had been no research conducted on the cheetah (Acinonyx jubatus)
population of the Northern Tuli Game Reserve (NOTUGRE), Botswana. My study aimed to
estimate the abundance and status of this population. I used two non-invasive techniques to
provide population estimates. Additionally, I determined the attitudes towards wildlife of rural
farmers living within and adjacent to the reserve and quantified the level of depredation on
livestock to understand the possible threats faced by cheetahs and other large predators on
communal farmlands.
6.1 Camera trapping: a monitoring technique for cheetahs
I used camera trapping as a method to estimate population density. The findings assisted in
developing camera trapping as a tool for deriving population estimates for cheetahs; a species
that occurs at low population densities (Caro 1994) and has relatively unpredictable movements
(see Chapter 3). Camera trapping is an affordable, repeatable and non-invasive method that can
be used to monitor cheetah populations where scent marking posts are known and accessible. The
method was refined from a method used in previous studies conducted in north-central Namibia
and the Thambazimbi district of the Limpopo Province in South Africa (Marker et al. 2008b;
Marnewick et al. 2008). When using scent marking posts, it is important to consider that there
may be variation in individual detectability (Otis et al. 1978). Male cheetahs may use scent
marking posts more frequently than females (Marnewick et al. 2006), therefore females may be
inadequately represented within the dataset both with regards to the number of captures and
identified individuals (Marker 2002; Marnewick et al. 2006; Marker et al. 2008b). Differences
in detection probabilities may be accounted for, where sample size permits, by incorporating sex-
specific encounter rates into Spatially Explicit Capture Recapture (SECR) models (O’Connell
et al. 2011). Nevertheless, a possible drawback of using cheetah-specific camera trapping sites
is that not all members of a predator guild can be simultaneously surveyed. Hence, the method
cannot be used for monitoring multiple species.
The recommended maximum number of days to maintain population closure when
conducting capture-recapture studies on large felids is 90 days (Karanth & Nichols 2002).
However, my study found a high latency to first cheetah detection (range: 9-85 days) and a
110
relatively low sample size (n = 18 independent cheetah capture events) after this initial 90
day survey. Extending the survey length (n = 130 days) increased the sample size (n = 31
independent cheetah captures) and the robustness of the density estimates. However, a
consequence of the long survey period is that population closure may have been violated
(Foster & Harmsen 2012). Furthermore, no significant differences were found in the density
estimates achieved. Increasing the surveyed area may reduce the effect of a high latency to
first detection as the spatial scale of my study area (±240 km²) may have been too small to
incorporate sufficient home ranges of cheetahs (see Chapter 4: distribution and home range).
Cheetahs have very large home ranges (Marker et al. 2008a) and I would therefore recommend
that when designing camera trap surveys for cheetahs, the area containing the trap array
should be large enough to incorporate sufficient home ranges. Maffei & Noss (2008)
recommend that the surveyed area should be at least three to four times the average home
range for the target species for that specific site. However, this requires a prior knowledge of the
home range size for cheetahs in that specific area, as the size of cheetahs’ home ranges varies
substantially between geographic locations due to differences in habitat type and prey density
(Caro 1994; Gros et al. 1996; Broomhall et al. 2003). Additionally, the spacing between camera
traps may be increased as to optimize trap spacing and accommodate for the relatively
large surveyed area required. Dillon & Kelly (2007) recommend at least two camera traps per
average home range. Furthermore, in place of the Adjacent Block method, the entire area
should be surveyed throughout the sampling period with traps set at half the density and moved
to their new location within the same area after half of the sampling period as suggested by
Foster & Harmsen (2012). This approach reduces the confounding effect of space and time
associated with the adjacent block method (Foster & Harmsen 2012). Deploying a single camera
trap unit at scent marking posts increased the total area that could be surveyed, although this
method can decrease capture probability due to variable trap effort from malfunctioning
equipment and missed individuals. Nonetheless, I feel that this approach was successful as
camera traps were set to either take short video clips or a burst of three images during a single
trigger event. Additionally, the scent marking posts functioned as a natural lure and thus
individuals would usually investigate the scent marking posts for a few minutes increasing the
chances of multiple images.
6.2 Citizen science in cheetah research
Citizen science, whereby volunteers assist with data collection, has become increasingly
important in ecological research (Silvertown 2009) as not only can a large amount of data be
quickly collected but it can also create awareness and a sense of conservation stewardship 111
(Silvertown 2009; Marnewick & Davies-Mostert 2012). Photographic survey methods employ
citizen scientists to collect information on the population ecology of rare, elusive and
individually identifiable species (Silvertown 2009). Cheetah specific photographic surveys
have been successfully implemented in a number of national parks, including the Kruger
National Park and the Kgalagadi Transfrontier Park in South Africa (Bowland & Mills 1994;
Knight 1999; Kemp & Mills 2005; Lindsey et al. 2009; Marnewick & Davies-Mostert 2012;
Marnewick et al. 2014). In this study I evaluated the use of a photographic survey to estimate
the minimum number and status of cheetahs in a private game reserve that receives fewer
visitors annually than National Parks. I also collected older photographs to assess differences in
numbers over time, and to provide estimates of age and relations of frequently sighted
individuals. Digital cameras are increasingly accessible and widely used, this is ideal for
photographic surveys as photographs retain the date and time at which they were taken. In
addition, photographs taken on smart phones or cameras with built in GPS’s can also record
the physical location of captures. Thus, data collected by the public can be validated by the
accompanying metadata (data specific to each photograph). A possible draw back to the
method is that the survey relies on incidental photographic records, thus photographs are
collected opportunistically and so the frequency of sightings cannot be used to assess space
use (e.g. habitat preference) and density as the location of sightings are invariably biased to
areas with higher tourism activity. Nonetheless, this method can be used as an ongoing
collection of photographs to monitor changes in population sizes. Additionally, photographic
surveys can provide baseline data on the number of cheetahs in a reserve, their distribution
and demography.
My study found that there was a marked difference observed in the sample effort between the
wet and dry season. More photographs were received from the cool dry season compared to
the wet season. The cool dry season marks the peak tourism period as well as the landscape
typically being more open which makes it easier to find cheetahs. It is therefore
recommended to carry out photographic surveys in the cool dry period to increase the total
sample size.
Camera traps have become increasingly accessible to the general public with a number of
private properties utilising these for recreational purposes (pers. obs.). My study demonstrates
the potential usefulness of camera traps as an additional tool for photographic surveys. The
majority (89%) of photographic entries submitted from South African farmlands were taken by
camera traps. This becomes particularly important where surveys are carried out outside of
112
National Parks or tourism-focused game reserves where cheetahs and other large carnivores
may be skittish and elusive.
In my study, I used the software program Adobe Photoshop Lightroom to manage all
photographic data, this software was developed for professional photographers to manage,
catalogue and edit large numbers of digital images. It is, therefore, ideal for camera trapping
and photographic survey data which typically have large volumes of photographs. I would
recommend making use of this program to assist with the identification process associated
with camera trapping and photographic surveys.
6.3 Conservation status of cheetahs in NOTUGRE
Although the cheetah has received a high amount of research and monitoring attention,
studies have been geographically biased to populations in Tanzania and Namibia and the
species is nevertheless still identified as a species of concern with high risk status (Ray et al.
2005). Botswana supports a significant number of free-roaming cheetahs and in an effort to
conserve the species, a ban on hunting cheetahs has been in place since 1968 (Klein 2007).
During this time, cheetahs could only be killed in defence of livestock. In 2000, a
memorandum was passed banning all killing of cheetahs (Klein 2007). However, the
repercussions for killing a cheetah, a 1000 Botswana Pula (BWP) fine (~US$ 100) or one
year in prison, may not be sufficient to discourage offenders (Klein 2007). Furthermore,
about half of the cheetahs in Botswana are estimated to range on unprotected farmlands
where habitats are undergoing degradation and the species may face persecution in retribution to
perceived or actual livestock depredation (Winterbach et al. 2014). Illegal trade of cheetahs,
particularly sub-adults and cubs, is also a cause of concern (Klein 2007). It is estimated
that between 50 and 60 cheetahs are illegally removed from Botswana annually (Klein 2007).
The survival of cheetahs in Botswana therefore appears precarious.
My findings suggest that cheetahs in NOTUGRE have a low population density and are
possibly undergoing a population decline. While a recovering lion population may contribute to
this decline (Laurenson 1994; Durant et al. 2004, Snyman et al. 2014), it might also be the result
of persecution as a result of conflict with livestock farmers outside NOTUGRE. Livestock
farmers whom I interviewed generally had a low tolerance for predators on farmlands.
Additionally, the results of my study suggest a high total livestock loss due to predators on
communal farmlands in comparison to other human-predator conflict studies (Graham et al.
2005). This may be attributed to the low abundance of natural prey, particularly larger species 113
which are believed to be mostly absent. The low wild prey abundance may be the result of
poaching but may also be attributed to habitat degradation on account of overstocking and poor
livestock management, thereby reducing the overall carrying capacity of the land (Klein
2007). A better understanding of the density of the natural prey base is, therefore, required.
Nonetheless, a lack of a proactive approach towards the raising of livestock was found to be
the primary cause for livestock depredation. Responsible farm management should be enforced
(Klein 2007), this requires farmers to regain self-responsibility for their livestock by
improving current livestock husbandry practices. The communities may benefit from an
incentivised program such as a locally managed communal insurance program (Mishra et al.
2003).
Although the high livestock loss to predators is a cause for concern as it may fuel human predator
conflict (Graham et al. 2005), a negative attitude towards predators in my study was not related
to livestock depredation. Farmers had an overall negative attitude towards conservation of large
predators on farmlands, but this was not related to economic losses, knowledge or other
demographic variables such as age or education, as was found in previous studies (Bath 1998;
Ericsson & Heberlein 2003; Røskaft et al. 2007; Selebatso et al. 2008). I suggest that attitudes
may be the product of social factors and human-human conflict rather than an economic loss
(Dickman 2010). Addressing human-human conflict and promoting an emotional affiliation
towards predators may, therefore, play a greater role in conflict resolution than reducing livestock
losses. Positive emotions and increasing appreciation for wildlife can be achieved by the
continued education of children (such as by the Children in the Wilderness program) and the
development of educational programs for local residents including exposing locals to positive
experiences with wildlife. Interest in the conservation of wildlife may be achieved by increasing
general awareness of the status of large predators but also by the potential financial returns
(Marker et al. 2003a).
Cheetahs do not hold any value to most farmers living on communal farmlands (see Chapter 5).
Implementing sustainable utilisation could give value to the wildlife and increase tolerance
towards predators and thereby their conservation (Klein 2007; Sifuna 2010). Consumptive use of
wildlife, however, requires accurate estimates of population densities in order to determine the
appropriate offtake (Klein 2007). Options for sustainable wildlife utilisation should be
investigated along with alternative livelihoods to livestock-keeping which may benefit
communities from coexisting with predators, including ecotourism, veld products, predator
friendly meat, and honey production (Klein 2007).
114
The absence of documented cheetah movement across the South African boundary is also of
concern, but may be a result of the limited number (n = 27) of photographic entries received
from South Africa. The persistence of genetic connectivity between sub-populations is essential
for the viability of the population in NOTUGRE, which will depend upon the level of
persecution as well as available movement corridors (free from human disturbances). It is
therefore imperative that conservation efforts incorporate the neighbouring farmlands
including those in neighbouring countries. However, policy and legislation varies across the
three states (Purchase et al. 2007). Botswana has listed the cheetah as protected and it cannot be
destroyed under any circumstances (Klein 2007). In Zimbabwe, the cheetah is protected but
can be killed with a permit from the Wildlife Management Authority (Purchase et al. 2007).
South Africa has complex legislation with each province providing its own regulations,
however the cheetah is protected to some degree and a permit is required to remove or kill
an animal (Purchase et al. 2007). Further research on conflict with cheetahs in neighbouring
countries is imperative and the creation, or maintenance of corridors to promote gene flow
should be incorporated in management considerations and policies. It is essential that
government authorities are involved in these decisions as they have the authority to
implement recommendations both at management and policy levels. Further research should
investigate whether links between these sub-populations exist, and identify potential
movement corridors between protected areas. This could be achieved by investing more
research efforts in neighbouring countries through an ongoing photographic census
particularly promoting the use of camera traps. Alternatively, movement of cheetahs could be
monitored by the use of satellite collars and genetic sampling could be used to determine
relatedness and hence the level of gene flow between South African, Zimbabwean and
Botswana cheetahs. Genetic sampling could further help determine whether there is more
than one genetically distinguishable population within the samples.
6.4 Conclusion
Cheetahs are challenging mammal species to study as they have large home ranges and occur at
low population densities (Gros 1998). This means that a population census has to be carried
out over a large area, and despite high sample effort, sample sizes are invariably low.
Demographically open populations, like NOTUGRE, are particularly challenging to monitor as
movement in and out of the study area is unknown and the population may straddle
different properties. Despite these challenges, I believe that my study produced valuable
information for the conservation and management of cheetahs in NOTUGRE, providing a
better understanding of local cheetah population size, status, and distribution in an area which 115
had previously not been researched; offering a baseline for future studies. My study further
provides valuable information on monitoring techniques for future research on cheetahs and
other large predators which occur at low population densities.
The cheetah is clearly a species of great concern with an elevated risk status and extensive
range loss, and therefore requires dedicated conservation efforts to prevent local extinction
(Ray et al. 2005). The cheetah is one of the most charismatic flagship species with substantial
economic and aesthetic value for the ecotourism industry and tourism-financed conservation
areas in Botswana. Furthermore, large predators, such as cheetahs, are key components of
ecosystems with flagship status and serve as an important umbrella species for the
conservation of biodiversity (Ray et al. 2005). Through the study of the predators of an
ecosystem the ecosystem as a whole is being studied, consequently the population status of a
top predator may serve as an indicator of overall ecosystem function and productivity (Packer et
al. 2003). Conservation actions directed towards large carnivore species, therefore, are
expected to have the greatest impact on overall ecosystem conservation (Buk & Marnewick
2010; Macdonald et al. 2010b).
116
Dusk settles over Mashatu on my last day of field work.
Photo: Eléanor Brassine
117
REFERENCES
AKAIKE, H. 1974. A new look at statistical model identification. IEEE Transactions on
Automatic Control 19: 716–723.
ALEXANDER, G.J. 1984. A preliminary investigation into the relationships between geology,
soils and vegetation in the eastern Tuli Block, Botswana. BSc (Hons) thesis. University of Natal.
Durban.
ANDRESEN, L., EVERATT, K.T. & SOMERS, M.J. 2014. Use of site occupancy models for
targeted monitoring of the cheetah. Journal of Zoology 292: 212–220.
BALINSKY B.I. 1962. Patterns of animal distribution on the African Continent (summing-up
talk). Annals of the Cape Provincial Museums. Natural History 2: 299-310.
BALME, G.A., HUNTER, L.T.B. & SLOTOW, R. 2009. Evaluating methods for counting
cryptic carnivores. The Journal of Wildlife Management 73(3): 433-441.
BALME, G.A., BATCHELOR, A., DE WORONIN BRITZ, N., SEYMOUR, G., GROVER,
M., HES, L., MACDONALD, D.W. & HUNTER, L.T.B. 2012. Reproductive success of female
leopards Panthera pardus: the importance of top-down processes. Mammal Review 43: 221–237.
BARTLAM-BROOKS, H.L.A., BONYONGO, M.C. & HARRIS, S. 2011. Will reconnecting
ecosystems allow long-term distance mammal migrations to resume? A case study of a zebra
Equus burchelli migration in Botswana. Oryx 45: 210-216.
BASHIR, S., DALY, B., DURANT, S.M., FÖRSTER, H., GRISHAM, J., MARKER, L.,
WILSON, K. & FRIEDMANN, Y. 2004. Global Cheetah (Acinonyx jubatus) Monitoring
Workshop. Final workshop report. Conservation Breeding Specialist Group (SSC/IUCN).
Endangered Wildlife Trust.
BATH, A.J. 1998. The role of human dimensions in wildlife resource research in wildlife
management. Ursus 10: 349–355.
BISSETT, C. 2004. The feeding ecology, habitat selection and hunting behaviour of re-
introduced cheetah on Kwandwe Private Game Reserve, Eastern Cape Province. MSc thesis.
Rhodes University. Grahamstown.
BISSETT, C. & BERNARD, R.T.F. 2007. Habitat selection and feeding ecology of the cheetah
(Acinonyx jubatus) in thicket vegetation: is the cheetah a savanna specialist? Journal of Zoology
271: 310–317.
BLAKE, J.G. & MOSQUERA, D. 2014. Camera trapping on and off trails in lowland forest of
eastern Ecuador: Does location matter? Mastozoología Neotropical 21(1): 17-26.
BORAH, J., SHARMA, T., DAS, D., RABHA, N., KAKATI, N., BASUMATARY, A.,
118
AHMED, M.F. & VATTAKAVEN, J. 2013. Abundance and density estimates forcommon
leopard Panthera pardus and clouded leopard Neofelis nebulosa in Manas National Park, Assam,
India. Oryx 48: 149–155.
BORCHERS, D. 2010. A non-technical overview of spatially explicit capture-recapture models.
Journal of Ornithology 152: 435-444.
BORCHERS, D.L. & EFFORD, M.G. 2008. Spatially explicit maximum likelihood methods for
capture-recapture studies. Biometrics 64: 377-385.
BÖRGER, L., FRANCONI, N., DE MICHELE, G., GANTZ, A., MESCHI, F., MANICA,
A., LOVARI, S. & COULSON, T. 2006. Effects of sampling regime on the mean and variance
of home range size estimates. Journal of Animal Ecology 75: 1393-1405.
BOTHMA, J. DU P. & WALKER, C. 1999. Larger Carnivores of the African Savannas. Van
Schaik Publishers, Pretoria.
BOWLAND, A.E. & MILLS, M.G.L. 1994. The 1990/1991 cheetah photographic survey.
Scientific Report 6/94. SANParks, Skukuza.
BROOMHALL, L.S., MILLS, M.G.L. & DU TOIT, J.T. 2003. Home range and habitat use by
cheetahs (Acinonyx jubatus) in the Kruger National Park. Journal of Zoology 261: 119–128.
BUK, K. & MARNEWICK, K. 2010. Cheetah conservation in South Africa. African Insight
39(4): 212-224.
BULTE, E.H. & RONDEAU, D. 2005. Why compensating for wildlife damages may be bad for
conservation. Journal of Wildlife Management 69: 14–19.
BURNHAM, K.P. & ANDERSON, D.R. 2002. Model selection and multimodel inference: a
practical information-theoretic approach. Second edition. Springer. New York.
BUTLER, J.R.A. 2000. The economic costs of wildlife predation on livestock in Gokwe
communal land, Zimbabwe. African Journal of Ecology 38: 23–30.
CARBONE, C., CHRISTIE, S., CONFORTI, K., COULDON, T., FRANKLIN, N., GINSBERG,
J.R., GRIFFITHS, M., HOLDEN, J., KAWANISHI, K., KINNAIRD, M., LAIDLAW, R.,
LYNAM, A., MACDONALD, D.W., MARTYR, D., MCDOUGAL, C., NATH, L., O’BRIEN,
T., SEIDENSTICKER, J., SMITH, D.J.L., SUNQUIST, M.,
TILSON, R. & WAN SHAHRUDDIN, W.N. 2001. The use of p h o t o g r a p h i c rates to
estimate densities of tigers and other cryptic mammals. Animal Conservation 4: 75-79.
CARO T.M. 1994. Cheetahs of the Serengeti Plains: group living in an asocial species. The
University of Chicago Press. Chicago.
CARO, T.M. & COLLINS, D.A. 1987. Male cheetah organization and territoriality. Ethology
74(1): 52-64.
119
CHATTHA, S.A., IQBAL, S., RASHEED, Z., RAZZAQ, A., HUSAIN, M. & ABBAS, M.N.
2013. Human-leopard conflict in Machiara National Park (MNP), Azad Jamu and Kashmir (AJ
and K), Pakistan. Jgiass 1: 17–21.
CLARKE, J. 2012. Save me from the Lion’s Mouth: Exposing Human-Wildlife Conflict in Africa.
Struik Nature. Cape Town.
COPPINGER, R., COPPINGER, L., LANGELOH, G., GETTLER, L. & LORENZ, J. 1988.
A decade of use of livestock guarding dogs. Proceedings of the Thirteenth Vertebrate Pest
Conference, University of Nebraska. Lincoln. 13: 209–214.
COZZI, G., BROEKHUIS, F., MCNUTT, J.W., SCHMID, B. 2013. Comparison of the effects of
artificial and natural barriers on large African carnivores: implications for interspecific
relationships and connectivity. Journal of Animal Ecology 82: 707-715.
DALTON, D.L., CHARRUAU, P., BOAST, L. & KOTZE, A. 2013. Social and genetic
population structure of free-ranging cheetah in Botswana: implications for conservation.
European Journal of Wildlife Research 59: 281–285.
DICKMAN, A.J. 2010. Complexities of conflict: the importance of considering social factors for
effectively resolving human-wildlife conflict. Animal Conservation 13: 458–466.
DILLON, A. & KELLY, M.J. 2007. Ocelot Leopardus pardalis in Belize: the impact of trap
spacing and distance moved on density estimates. Oryx 41: 469–477.
DURANT, S.M. 1998. Competition refuge and coexistence: an example from Serengeti
carnivores. Journal of Animal Ecology 67: 370-386.
DURANT, S.M., KELLY, M. & CARO, T.M. 2004. Factors affecting life and death in Serengeti
cheetahs: environment, age, and sociality. Behavioural Ecology 15(1): 11- 22.
DURANT, S.M., BASHIR, S., MADDOX, T. & LAURENSON, M.K. 2007. Relating long- term
studies to conservation practice: the case of the Serengeti Cheetah Project.
Conservation biology : the journal of the Society for Conservation Biology 21: 602–11.
DURANT, S.M., DICKMAN, A.J. MADDOX, T., WAWERU, M.N., CARO, T.,
PETTORELLI, N., MACDONALD, D.W. & LOVERIDGE, A.J. 2010. Past, present, and future
of cheetahs in Tanzania: their behavioural ecology and conservation. In Macdonald, D.W. &
Loveridge, A.J. Editors. Biology and Conservation of Wild Felids. Oxford University Press.
Oxford. P. 373-382.
DWNP (The Department Of Wildlife and National Parks). 2009. National Conservation Action
plan for Cheetahs and African Wild Dog in Botswana.
120
DWNP (The Department Of Wildlife and National Parks). 2012. Aerial census of animals in
Botswana: Dry season.
DYER, S. 2012. Population size, demography and spatial ecology of cheetahs in the Timbavati
Private Nature Reserve, South Africa. MSc thesis. Rhodes University. Grahamstown.
EATON, R.L. 1970. Group interactions, spacing and territoriality in cheetahs. Zeitschrift für
Tierpsychologie 27(4): 482–491.
EFFORD, M. 2004. Density estimation in live-trapping studies. Oikos 106: 598-610.
EFFORD, M.G., DAWSON, D.K. & ROBBINS, C.S. 2004. DENSITY: software for analysing
capture – recapture data from passive detector arrays. Animal Biodiversity and Conservation 27:
217–228.
ELLIOT, N.B., VALEIX, M., MACDONALD, D.W. & LOVERIDGE, A.J. 2014. Social
relationships affect dispersal timing revealing a delayed infanticide in African lions.
Oikos 123: 1049–1056.
ERICSSON, G. & HEBERLEIN, T.A. 2003. Attitudes of hunters, locals, and the general public
in Sweden now that the wolves are back. Biological Conservation 111: 149–159.
FERREIRA, S.M.& HOFMEYR, M. 2014. Managing charismatic carnivores in small areas: large
felids in South Africa. South African Journal of Wildlife Research 44: 32–42.
FODDY, W.H. 1994. Constructing Questions for Interviews and Questionnaires: Theory and
Practice in Social Research. Cambridge University Press. Cambridge.
FOSTER, R.J. & HARMSEN, B.J. 2012. A critique of density estimation from camera-trap data.
The Journal of Wildlife Management 76(2): 224–236.
FRECKLETON, R. P. 2011. Dealing with collinearity in behavioural and ecological data: model
averaging and the problems of measurement error. Behavioural Ecology and Sociobiology 65:
91–101.
GARDNER, B., REPPUCCI, J., LUCHERINI, M. & ROYLE, J.A. 2010. Spatially explicit
inference for open populations: estimating demographic parameters from camera-trap studies.
Ecology 91(11): 3376-3383.
GERBER, B.D., KARPANTY, S.M.& KELLY, M.J. 2012. Evaluating the potential biases in
carnivore capture-recapture studies associated with the use of lures and varying density
estimation techniques using photographic-sampling data of the Malagasy civet.
Population Ecology 54: 43-54.
GLIEM, J.A. & GLIEM, R.R. 2003. Calculating, Interpreting, and Reporting Cronbach’s Alpha
Reliability Coefficient for Likert-Type Scales. Midwest Research to Practice Conference in Adult,
Continuing, and Community Education. pp. 82–88.
GONZALEZ, A., NOVARDO, A., FUNES, M., PAILACURA, O., BOLGERI, M.J. & 121
WALKER, S. 2012. Mixed-breed guarding dogs reduce conflict between goat herders and native
carnivores in Patagonia. Human-Wildife Interactions 6: 327–334.
GOPALASWAMY, A.M., ROYLE, J.A., HINES, J.E., SINGH, P., JATHANNA, D.,
KUMAR, N.S. & KARANTH, K.U. 2012. Program SPACECAP: software for estimating animal
density using spatially explicit capture-recapture models. Methods in Ecology and Evolution 3:
1067-1072.
GOPALASWAMY, A.M., ROYLE, A.J., HINES, J.E., SINGH, P., JATHANNA, D.,
KUMAR, N.S. & KARANTH, K.U. 2013. Package ‘SPACECAP’ A program to estimate animal
abundance and density using Bayesian spatially explicit capture- recapture models.
GÖTZE, A.R., CILLIERS, S.S., BEZUIDENHOUT, H. & KELLNER, K. 2008. Analysis of the
vegetation of the sandstone ridges (ib land type) of the north-eastern parts of the Mapungubwe
National Park, Limpopo Province, South Africa. Koedoe: African Protected Area Conservation
and Science 50: 72–81.
GRAHAM, K., BECKERMAN, A.P. & THIRGOOD, S. 2005. Human–predator–prey conflicts:
ecological correlates, prey losses and patterns of management. Biological Conservation 122:
159–171.
GRANT, T. 2012. Leopard population density, home range size and movement patterns in a
mixed landuse area of the Mangwe district of Zimbabwe. MSc thesis, Rhodes University.
Grahamstown.
GRAY, T.N.E. & PRUM, S. 2012. Leopard density in post-conflict landscape, Cambodia:
Evidence from spatially explicit capture-recapture. The Journal of Wildlife Management 76(1):
163–169.
GROS, P.M. 1998. Status of the cheetah Acinonyx jubatus in Kenya: a field-interview assessment.
Biological Conservation 85: 137-149.
GROS, P.M. 2000. Status of the cheetah in Tanzania in the mid 1990’s. Journal of East African
Natural History 89: 85–100.
GROS, P.M. 2002. The status and conservation of the cheetah Acinonyx jubatus in Tanzania.
Biological Conservation 106: 177–185.
GROS, P.M., KELLY, M. J. & CARO, T. M. 1996. Estimating carnivore densities for
conservation purposes: indirect methods compared to baseline demographic data. Oikos 77: 197–
206.
HALL-MARTIN, A.J. & BOTHA, B.P. 1980. A note on feeding habits, ectoparasites and
measurements of black-backed jackal Canis mesomelas from Addo Elephant National Park.
Koedoe 23: 157-162.
HARRIS, S., CRESSWELL, W.J., FORDE, P.G., TREWHELLA, W.J., WOOLLARD, T. & 122
WRAY, S. 1990. Home-range analysis using radio-tracking data: a review of the problems and
techniques particularly as applied to the study of mammals. Mammal Review 20: 97-123.
HAYWARD, M.W. & SLOTOW, R. 2009. Temporal partitioning of activity in large African
carnivores: tests of multiple hypotheses. South African Journal of Wildlife Research 39(2): 109-
125.
HEBERLEIN, T.A. 2012. Navigating Environmental Attitudes. Oxford University Press.
New York.
HEILBRUN, R.D., SILVY, N.J., TEWES, M.E. & PETERSON, M.J. 2003. Using automatically
triggered cameras to individually identify bobcats. Wildlife Society Bulletin 31: 748–755.
HENSCHEL, P. & RAY, J. 2003. Leopards in African Rainforests: Survey and monitoring
techniques. WCS Global Carnivore Program.
HOLMERN, T., NYAHONGO, J. & RØSKAFT, E. 2007. Livestock loss caused by predators
outside the Serengeti National Park, Tanzania. Biological Conservation 135: 518–526.
HOLMERN, T. & RØSKAFT, E. 2013. The poultry thief: Subsistence farmers’ perceptions of
depredation outside the Serengeti National Park, Tanzania. African Journal of Ecology 52: 3334–
342.
HOLMERN, T., NYAHONGO, J. & RØSKAFT, E. 2007. Livestock loss caused by predators
outside the Serengeti National Park, Tanzania. Biological Conservation 135(4): 518– 526.
HOUSER, A., SOMERS, M. J. & BOAST, L.K. 2009. Home range use of free-ranging cheetah
on farm and conservation land in Botswana. South African Journal of Wildlife Research 39(1):
11-22.
HUDENKO, H.W. 2012. Exploring the influence of emotion on human decision making in
human–wildlife conflict. Human Dimensions of Wildlife: An International Journal 17: 16–28.
HUNTER, L. 1998. Early post-release movements and behaviour of reintroduced cheetahs and
lions, and technical considerations in large carnivore restoration. Proceedings of a Symposium
on Cheetahs as Game Ranch Animals. Ondersterpoort: 72-82.
IUCN (International Union for Conservation of Nature). 2013. IUCN Red List of Threatened
Species. Version 2013.2. Available at: http://www.iucnredlist.org. Accessed: 10 April 2014.
JACKSON, T.P., MOSOJANE, S., FERREIRA, S.M. & VAN AARDE, R.J. 2008. Solutions for
elephant Loxodonta africana crop raiding in northern Botswana: moving away from symptomatic
approaches. Oryx 42: 83–91.
JACKSON, R.M., ROE, J.D., WANGCHUK, R. & HUNTER, D.O. 2010. Estimating snow
leopard population abundance using photography and capture-recapture techniques. Wildlife
Society Bulletin 34(3): 772-781.
123
JACKSON, C.R., MCNUTT, J.W. & APPS, P.J. 2012. Managing the ranging behaviour of
African wild dogs (Lycaon pictus) using translocated scent marks. Wildlife Research 39: 31–34.
JENNESS, J. 2012. Repeating shapes for ArcGIS. Jenness Enterprises. Available at:
http://www.jennessent.com/arcgis/repeat_shapes.htm. Accessed: 12 January 2014.
JOUBERT, S.C.J. 1984. Naledi Game Reserve: Report on an ecological survey of the Eastern
Tuli Game Reserve. Unpublished report.
KALLE, R., RAMESH, T., QURESHI, Q. & SANKAR, K. 2011. Density of tiger and leopard in
a tropical deciduous forest of Mudumalai Tiger Reserve, southern India, as estimated using
photographic capture–recapture sampling. Acta Theriologica 56: 335– 342.
KARANTH, K.U. 1995. Estimating tiger Panthera tigris populations from camera-trap data
using capture-recapture models. Biological Conservation 71: 333-338.
KARANTH, K.U. & NICHOLS, J.D. 1998. Estimation of tiger densities in India using
photographic captures and recaptures. Ecology 79(8): 2852-2862.
KARANTH, K.U. & NICHOLS, J.D. 2002. Monitoring tigers and their prey: a manual for
researchers, managers and conservationists in tropical Asia. Centre for Wildlife Studies.
Bangalore, India.
KARANTH, K.U., FUNSTON, P. & SANDERSON, E. MACDONALD, D.W. &
LOVERIDGE, A.J. 2010. Many ways of skinning a cat: tools and techniques for studying wild
felids. In Macdonald, D.W. & Loveridge, A.J. Editors. Biology and Conservation of Wild Felids.
Oxford University Press. Oxford. p. 197-216.
KELLY, M.J. 2001. Computer-aided photography matching in studies using individual
identification: an example from Serengeti cheetahs. Journal of Mammalogy 82(2): 440- 449.
KELLY, M. J. & DURANT, S. M. 2000. Viability of the Serengeti Cheetah Population.
Conservation Biology 14: 786–797.
KELLY, M.J., LAURENSON, M.K., FITZGIBBON, C.D., COLLINS, D.A., DURANT, S.M.,
FRAME, G.W., BERTRAM, B.C.R. & CARO, T.M. 1998. Demography of the Serengeti cheetah
(Acinonyx jubatus) population: the first 25 years. Journal of Zoology 244: 473-488.
KELLY, M.J., NOSS, A.J., DI BITETTI, M.S., MAFFEI, L., ARISPE, R.L., PAVIOLO, A.,
DE ANGELO, C.D. & DI BLANCO Y.E. 2008. Estimating Puma densities from camera trapping
across three study sites: Bolivia, Argentina, and Belize. Journal of Mammalogy 89(2): 408-418.
KEMP, L.V. & MILLS, M.G.L. 2005. The 4th wild dog and 2nd cheetah photographic census in
the greater Kruger region, September 2004 – April 2005. Unpublished report. Endangered
Wildlife Trust, Johannesburg.
KGATHI, D.L., MMOPELWA, G., MASHABE, B. & MOSEPELE, K. 2012. Livestock
predation, household adaptation and compensation policy: a case study of Shorobe Village in 124
northern Botswana. Agrekon: Agricultural Economics Research, Policy and Practice in Southern
Africa 51: 22–37.
KLEIN, R. 2007. Status Report for the Cheetah in Botswana. Published by Cheetah Conservation
Botswana for the IUCN’s Southern Africa Regional Status Report. 14–21.
KLEIN, R. 2013. An assessment of human carnivore conflict in the Kalahari region of Botswana.
MSc thesis. Rhodes University. Grahamstown.
KNIGHT, A. 1999. Cheetah numbers in a changing environment: Kalahari Gemsbok National
Park. Endangered Wildlife Trust, Johannesburg.
LANDA, A., GUDVANGEN, K., SWENSON, J E. & RØSKAFT, E. 1999. Factors associated
with wolverine Gulo gulo predation on domestic sheep. Journal of Animal Ecology 36: 963–973.
LAURENSON, M.K. 1994. High juvenile mortality in cheetahs (Acinonyx jubatus) and its
consequences for maternal care. Journal of Zoology 234: 387-408.
LAURENSON, M.K., WIELEBNOWSKI, N. & CARO, T.M. 1995. Extrinsic factors and
juvenile mortality in cheetahs. Conservation Biology 9: 1329–1331.
LIND, P. 1974. Shashe Limpopo Ranger Report 1973/10974. Unpublished report.
LINDSEY, P., DU TOIT, J.T. & MILLS, M.G.L. 2004. The distribution and population status of
African wild dogs (Lycaon pictus) outside protected areas in South Africa. South African Journal
of Wildlife Research 34(2): 143-151.
LINDSEY, P.A., HAVEMANN, C.P., LINES, R., PALAZY, L., PRICE, A.E., RETIEF,
T.A., RHEBERGEN, T. & VAN DER WAAL, C. 2013. Determinants of persistence and
tolerance of carnivores on Namibian ranches: implications for conservation on southern African
private lands. PLoS ONE 8(1): e52458.
LINNELL, J.D.C., SWENSON, J.E. & ANDERSEN, R. 2001. Predators and people: conservation
of large carnivores is possible at high human densities if management policy is favourable.
Animal Conservation 4: 345–349.
LONG, R.A., MACKAY, P., ZIELINSKI, W.J. & RAY, J.C. 2008. Noninvasive Survey
Methods for Carnivores. Island Press, Washington.
MACDONALD, D.W., LOVERIDGE, A.J. & NOWELL, K. 2010a. Dramatis personae: an
introduction to the wild felids. In Macdonald, D.W. & Loveridge, A.J. Editors. Biology and
Conservation of Wild Felids. Oxford University Press. Oxford. p. 3-58.
MACDONALD, D.W., LOVERIDGE, A.J. & RABINOWITZ, A. 2010b. Felid futures: crossing
disciplines, borders, and generations. In Macdonald, D.W. & Loveridge, A.J. Editors. Biology
and Conservation of Wild Felids. Oxford University Press. Oxford. p. 599-649.
125
MADDOCK, A.H. & MILLS, M.G.L. 1994. Population characteristics of African wild dogs
lycaon pictus in the Eastern Transvaal Lowveld, South Africa, as revealed through photographic
records. Biological Conservation 67: 57-62.
MAFFEI, L. & NOSS, A.J. 2008. How small is too small? Camera trap survey areas and density
estimates for ocelots in the Bolivian Chaco. Biotropica 40(1): 71–75.
MAFFEI, L., NOSS, A.J., CUELLAR, E. & RUMIZ, D.I. 2005. Ocelot (Felis pardalis) population
densities, activity, and ranging behaviour in the dry forests of eastern Bolivia: data from camera
trapping. Journal of Tropical Ecology 21: 349-353.
MARKER, L. 1998. The current status of the cheetah (Acinonyx jubatus). Preceedings of a
Symposium on Cheetahs as Game Ranch Animals, Onderstepoort.
MARKER, L. 2000. Aspects of the ecology of the cheetah (Acinonyx jubatus) on the north central
Namibian farmlands. Namibian Scientific Society Journal 48: 40-48.
MARKER, L. L. 2002. Aspects of cheetah (Acinonyx jubatus) biology, ecology and conservation
strategies on Namibian Farmlands. PhD thesis. University of Oxford. Oxford.
MARKER, L.L., MILLS, M.G.L. & MACDONALD, D.W. 2003a. Factors influencing
perceptions of conflict and tolerance toward cheetahs on namibian farmlands.
Conservation Biology 17: 1290–1298.
MARKER, L.L., DICKMAN, A.J., JEO, R.M., MILLS, M.G.L. & MACDONALD, D.W.
2003b. Demography of the Namibian cheetah, Acinonyx jubatus jubatus. Biological
Conservation 114: 413–425.
MARKER, L.L., DICKMAN, A.J., MILLS, M.G.L. & MACDONALD, D.W. 2003c.
Aspects of the management of cheetahs, Acinonyx jubatus jubatus, trapped on Namibian
farmlands. Biological Conservation 114: 401–412.
MARKER, L.L., MUNTIFERING, J.R., DICKMAN, A.J., MILLS, M.G.L. &
MACDONALD, D. W. 2003d. Quantifying prey preferences of free-ranging Namibian cheetahs.
South African Journal of Wildlife Research 33: 43–53.
MARKER, L., DICKMAN, A. & SCHUMANN, M. 2005. Using livestock guarding dogs as a
conflict resolution strategy on Namibian farms. Carnivore Damage Prevention News: 28–32.
MARKER, L., DICKMAN, A., WILKINSON, C., SCHUMANN, B. & FABIANO, E. 2007.
The Namibian Cheetah : Status Report. CAT News: 4–13.
MARKER, L.L., DICKMAN, A.J., MILLS, M.G.L., JEO, R.M. & MACDONALD, D.W.
2008a. Spatial ecology of cheetahs on north-central Namibian farmlands. Journal of Zoology
274: 226-238.
MARKER, L.L., FABIANO, E. & NGHIKEMBUA, M. 2008b. The use of remote camera traps
to estimate density of free-ranging cheetahs in North-Central Namibia. CAT News 49: 22-24. 126
MARKER, L., DICKMAN, A.J., MILLS, M.G.L. MACDONALD, D.W. & LOVERIDGE, A.
2010. Cheetahs and ranchers in Namibia: a case study. In Macdonald, D.W. & Loveridge, A.J.
Editors. Biology and Conservation of Wild Felids. Oxford University Press. Oxford. P. 353-372.
MARNEWICK, K. & CILLIERS, D. 2006. Range use of two coalitions of male cheetahs,
Acinonyx jubatus, in the Thabazimbi district of the Limpopo Province, South Africa. South
African Journal of Wildlife Research 36(2): 147–151.
MARNEWICK, K. & DAVIES-MOSTERT, H.T. 2012. Kruger National Park 2008/2009 5th
Wild Dog and 2nd Cheetah Photographic Census. Report to the Endangered Wildlife Trust and
SANParks
MARNEWICK, K.A., BOTHMA, J. du P. & VERDOORN, G.H. 2006. Using camera- trapping
to investigate the use of a tree as a scent-marking post by cheetahs in the Thabazimbi district.
South African Journal of Wildlife Research 36(2): 139-145.
MARNEWICK, K., BECKHELLING, A., CILLIERS, D., LANE, E., MILLS, G., HERING, K.,
CALDWELL, P., HALL, R. & MEINTJES, S. 2007. The Status of the Cheetah in South Africa.
CAT News: 22–31.
MARNEWICK, K., FUNSTON, P.J. & KARANTH, K.U. 2008. Evaluating camera trapping as
a method for estimating cheetah abundance in ranching areas. South African Journal of Wildlife
Research 38(1): 59-65.
MARNEWICK, K., FERREIRA, S.M., GRANGE, S., WATERMEYER, J., MAPUTLA, N.
& DAVIES-MOSTERT, H.T. 2014. Evaluating the status of and African wild dogs Lycaon pictus
and cheetahs Acinonyx jubatus through tourist-based photographic surveys in the Kruger National
Park. PloS one 9: 1–8.
MCKENZIE, A.A. 1990. Co-operative hunting in the black-backed jackal Canis mesomelas
Schreber. PhD thesis. University of Pretoria, Pretoria.
MILLS, M.G.L. 1989. Cheetah and wild dog research in the Kruger National Park in 1988. A
progress report. Quagga: 5-6.
MISHRA, C., ALLEN, P., MCCARTHY, T., MADHUSUDAN, M.D., BAYARJARGAL, A. &
PRINS, H.H.T. 2003. The role of incentive programs in conserving the snow leopard.
Conservation Biology 17: 1512–1520.
MONDAL, K., GUPTA, S., BHATTACHARJEE, S., QURESHI, Q. & SANKAR, K. 2012.
Response of leopards to re-introduced tigers in Sariska Tiger Reserve, Western India.
International Journal of Biodiversity and Conservation 4(5): 228–236.
MYERS, N. 1975. The cheetah, Acinonyx jubatus, in Africa. IUCN Monograph No. 4,
International Union for Nature and Natural Resources, Morges, Switzerland. 127
NCHUNGA, M.L. 1978. A study of the potential for the commercial use of wildlife in the North-
eastern Tuli Block. MSc thesis. Wildlife and Fisheries Sciences, Graduate College of Texas A &
M University.
NEGRÕES, N., SARMENTO, P., CRUZ, J., EIRA, C., REVILLA, E., FONSECA, C.,
SOLLMANN, R., TÕRRES, N.M., FURTADO, M.M., JÁCOMO, A.T.A. & SILVEIRA, L.
2010. Use of camera-trapping to estimate puma density and influencing factors in central Brazil.
Wildlife Management 74(6): 1195-1203.
NEGRÕES, N., SOLLMANN, R., FONSECA, C., JÁCOMO, A.T.A., REVILLA, E. &
SILVEIRA, L. 2012. One or two cameras per station? Monitoring jaguars and other mammals in
the Amazon. Ecological Research 27: 639-648.
NICHOLS, J.D. 1992. Capture-recapture models. Biological Science 42(2): 94-101. NOSS, A.J.,
GARDNER, B., MAFFEI, L., CUÉLLAR, E., MONTAÑO, R., ROMERO-
MUNOZ, A., SOLLMAN, R. & O’CONNELL, A.F. 2012. Comparison of density estimation
methods for mammal populations with camera traps in the Kaa-Iya del Gran Chaco landscape.
Animal Conservation 15: 527–535.
NOSS, A., POLISAR, J., MAFFEI, L., GARCIA, R. & SILVER, S. 2013. Evaluating jaguar
densities with camera traps. Jaguar Conservation Program, Wildlife Conservation Society. New
York.
O’BRIEN, T.G. & KINNAIRD, M.F. 2011. Density estimation of sympatric carnivores using
spatially explicit capture-recapture methods and standard trapping grid. Ecological Applications
21(8): 2908-2916.
O’BRIEN, T.G., KINNAIRD, M.F. & WIBISONO, H.T. 2003. Crouching tigers, hidden prey:
Sumatran tiger and prey populations in a tropical forest landscape. Animal Conservation 6: 131
139.
O’CONNELL, A.F., NICHOLS, J.D. & KARANTH, K.U. 2011. Camera Traps in Animal
Ecology Methods and Analyses. Springer, Tokyo.
OGADA, M.O., WOODROFFE, R., OGUGE, N.O. & FRANK, L.G. 2003. Limiting
Depredation by African Carnivores: the Role of Livestock Husbandry. Conservation Biology
17(6): 1521–1530.
OGARA, W.O., GITAHI, N.J., ANDANJE, S.A., OGUGE, N., NDUATI, D.W., &
MAINGA, A.O. 2010. Determination of carnivores prey base by scat analysis in Samburu
community group ranches in Kenya. African Journal of Environmental Science and Technology
4(8): 540–546.
OLSON, K. 2010. An examination of questionnaire evaluation by expert reviewers.
128
Sociology Department, Faculty Publications. Paper 136.
http://digitalcommons.unl.edu/sociologyfacpub/136. Accessed: 1 October 2014.
OTIS, D.L., BURNHAM, K.P., WHITE, G.C. & ANDERSON, D.R. 1978. Statistical inference
from capture data on closed animal populations. Wildlife Monographs 62: 3- 135.
PACKER, C., HOLT, R.D., HUDSON, P.J., LAFFERTY, K.D. & DOBSON, A.P. 2003.
Keeping the herds healthy and alert: implications of predator control for infectious disease.
Ecology Letters 6: 797–802.
PARKER, D.M., WHITTINGTON-JONES, B.M., BERNARD, R.T.F. & DAVIES- MOSTERT,
H.T. 2014. Attitudes of rural communities toward dispersing african wild dogs in South Africa.
Human Dimensions of Wildlife 19: 512–522.
PATTERSON, B.D., KASIKI, S.M., SELEMPO, E. & KAYS, R.W. 2004. Livestock predation
by lions (Panthera leo) and other carnivores on ranches neighboring Tsavo National Park, Kenya.
Biological Conservation 119: 507–516.
PETTIFER, H.L. 1981. Aspects on the ecology of cheetahs (Acinonyx jubatus) on the
Suikerbosrand Nature Reserve. In: Proceedings of the First World Furbearer Conference. (eds)
J.A. Chapman & D. Punsley, pp. 1121-1142. University of Maryland, Frostburg. Virginia.
PETTORELLI, N., HILBORN, A., BROEKHUIS, F. & DURANT, S.M. 2009. Exploring habitat
use by cheetahs using ecological niche factor analysis. Journal of Zoology 277: 141–148.
POLISAR, J., MAXIT, I., SCOGNAMILLO, D., FARRELL, L., SUNQUIST, M.E. &
EISENBERG, J.F. 2003. Jaguars, pumas, their prey base, and cattle ranching: ecological
interpretations of a management problem. Biological Conservation 109: 297–310.
PRUGH, L R., STONER, C.J., EPPS, C.W., BEAN, W.T., RIPPLE, W.J., LALIBERTE, A.S. &
BRASHARES, J.S. 2009. The rise of the mesopredator. BioScience 59: 779– 791.
PURCHASE, G.K. & DU TOIT, J.T. 2000. The use of space and prey by cheetah in Matusadona
National Park, Zimbabwe. South African Journal of Wildlife Research 30(4): 139-144.
PURCHASE, G., MARKER, L., MARNEWICK, K., KLEIN, R. & WILLIAMS, S. 2007.
Regional Assessment of the Status, Distribution and Conservation Needs of Cheetahs in Southern
Africa. CAT News: 44–46.
RAMAKRISHNAN, U., COSS, R.G. & PELKEY, N.W. 1999. Tiger decline caused by the
reduction of large ungulate prey: evidence from a study of leopard diets in southern India.
Biological Conservation 89: 113–120.
RAY, J.C., HUNTER, L. & ZIGOURIS, J. 2005. Setting conservation and research priorities for
larger African carnivores. Wildlife Conservation Society, New York, USA.
RØSKAFT, E., HÄNDEL, B., BJERKE, T. & KALTENBORN, B.P. 2007. Human attitudes
towards large carnivores in Norway. Wildlife Biology 13: 172–185. 129
ROWE, R.J. 2009. Environmental and geometric drivers of small mammal diversity along
elevational gradients in Utah. Ecography 32: 411–422.
ROYLE, J.D., KARANTH, K.U., GOPALASWAMY, A.M. & KUMAR, N.S. 2009a. Bayesian
inference in camera trapping studies for a class of special capture-recapture models. Ecological
Society of America 90(11): 3233-3244.
ROYLE, J.D., NICHOLS, J.D., KARANTH, K.U. & GOPALASWAMY, A.M. 2009b. A
hierarchical model for estimating density in camera-trap studies. Journal of Applied Ecology 46:
118-127.
SANTOS, J.R.A. 1999. Cronbach’s Alpha: a tool for assessing the reliability of scales. Extension
Information Technology 37(2): 1-4.
SCHIESS-MEIER, M., RAMSAUER, S., GABANAPELO, T. & KÖNIG, B. 2007. Livestock
predation—insights from problem animal control registers in Botswana. Journal of Wildlife
Management 71: 1267–1274.
SEAMAN, D E. & POWELL, R.A. 1996. An evaluation of the accuracy of kernel density
estimators for home range analysis. Ecology 77: 2075–2085.
SELEBATSO, M., MOE, S.R. & SWENSON, J.E. 2008. Do farmers support cheetah Acinonyx
jubatus conservation in Botswana despite livestock depredation? Oryx 42(3): 430-436.
SELIER, S. J. 2007. The social structure, distribution and demographic status of the African
elephant population in the Central Limpopo River Valley of Botswana, Zimbabwe and South
Africa. MSc thesis. University of Pretoria. Pretoria.
SIFUNA, N. 2010. Wildlife damage and its impact on public attitudes towards conservation: a
comparative study of Kenya and Botswana, with particular reference to Kenya’s Laikipia Region
and Botswana's Okavango Delta Region. Journal of Asian and African Studies 45: 274–296.
SILVER, S. 2004. Assessing jaguar abundance using remotely triggered cameras. Jaguar
Conservation Program. Wildlife Conservation Society.
SILVERTOWN, J. 2009. A new dawn for citizen science. Trends in ecology & evolution 24:
467–471
SKINNER, J.D. & CHIMIMBA, C.T. 2005. The Mammals of the Southern African Subregion.
Cambridge University Press, Cambridge.
SNYMAN, A., JACKSON, C.R. & FUNSTON, P.J. 2014. The effect of alternative forms of
hunting on the social organization of two small populations of lions Panthera leo in southern
Africa. Oryx: 1–7.
SOISALO, M.K. & CAVALCANTI, S.M.C. 2006. Estimating the density of a jaguar population
in the Brazilian Pantanal using camera-traps and capture-recapture sampling in combination with
GPS radio-telemetry. Biological Conservation 129: 487-496. 130
SOLLMANN, R., FURTADO, M.M., GARDNER, B., HOFER, H., JÁCOMO, A.T.A.,
TÔRRES, N.M. & SILVEIRA, L. 2011. Improving density estimates for elusive carnivores:
accounting for sex-specific detection and movements using spatial capture– recapture models for
jaguars in central Brazil. Biological Conservation 144: 1017–1024.
SOSO, S.B., KOZIEL, J.A., JOHNSON, A., LEE, Y.J. & FAIRBANKS, W.S. 2014. Analytical
methods for chemical and sensory characterization of scent-marking in large wild mammals: a
review. Sensors 14: 4428-4465.
STANLEY, T.R. & BURNHAM, K.P. 1999. A closure test for time–specific capture- recapture
data. Environmental and Ecological Statistics 6: 197-209.
STEYN, T. 2004. Northern Tuli Game Reserve: Memories of the Founding of a Major Private
Game Reserve. E & R Steyn. Kelvin, Gauteng.
SUN, C.C., FULLER, A.K. & ROYLE, J.A. 2014. Trap configuration and spacing influences
parameter estimates in spatial capture-recapture models. PloS ONE 9(2): 1-9.
THOMPSON, S.K., WHITE, G.C. & GOWAN, C. 1998. Monitoring vertebrate populations.
Academic Press, New York, United States of America.
THORN, M., GREEN, M., DALERUM, F., BATEMAN, P.W. & SCOTT, D.M. 2012. What
drives human – carnivore conflict in the North West Province of South Africa ?
Biological Conservation 150: 23–32.
THORN, M., GREEN, M., MARNEWICK, K. & SCOTT, D.M. 2014. Determinants of attitudes
to carnivores: implications for mitigating human–carnivore conflict on South African farmland.
Oryx: 1–8.
TOBLER, M.W. & POWELL, G.V N. 2013. Estimating jaguar densities with camera traps:
problems with current designs and recommendations for future studies. Biological Conservation
159: 109–118.
TOBLER, M.W., CARRILLO-PERCASTEGUI, S.E., LEITE PITMAN, R., MARES, R. &
POWELL, G. 2008. An evaluation of camera traps for inventorying large- and medium- sized
terrestrial rainforest mammals. Animal Conservation 11: 169–178.
TOBLER, M.W., CARRILLO-PERCASTEGUI, S.E., ZÚÑIGA HARTLEY, A. & POWELL,
G.V.N. 2013. High jaguar densities and large population sizes in the core habitat of the
southwestern Amazon. Biological Conservation 159: 375–381.
TROLLE, M. & KERY, M. 2003. Estimation of ocelot density in the Pantanal using capture-
recapture analysis of camera-trapping data. Journal of Mammalogy, 84(2): 607-614.
WALKER, C.H. 1971. Field impressions: Tuli Circle Rhodesia, Northern Tuli Block Botswana.
Unpublished Report.
131
WALPOLE, M.J. & GOODWIN, H.J. 2001. Local attitudes towards conservation and tourism
around Komodo National Park, Indonesia. Environmental Conservation 28: 160–166.
WATERMEYER, J. 2012. Anthropogenic threats to resident and dispersing African wild dogs
west and south of the Kruger National Park, South Africa. MSc. thesis. Rhodes University.
Grahamstown.
WHITE, G.C., ANDERSON, D.R., BURNHAM, K.P. & OTIS, D.L. 1982. Capture- recapture
and removal methods for sampling closed populations. Los Alamos National Laboratory
Publications LA-8787-NERP. Los Alamos, New Mexico, United States of America.
WHITE, P.C.L., JENNINGS, N.V., RENWICK, A.R. & BARKER, N.H.L. 2005. Questionnaires
in ecology: a review of past use and recommendations for best practice. Journal of Applied
Ecology 42: 421–430.
WINTERBACH, H.E.K., WINTERBACH, C.W. & SOMERS, M.J. 2014. Landscape suitability
in Botswana for the conservation of its six large African carnivores. PloS one 9: 1–12.
WOODROFFE, R. 2000. Predators and people: using human densities to interpret declines of
large carnivores. Animal Conservation 3: 165–173.
WOODROFFE, R. & GINSBERG, J.R. 1998. Edge Effects and the Extinction of Populations
Inside Protected Areas. Science 280: 2126–2128.
WOODROFFE, R. & GINSBERG, J.R. 1999. Conserving the African wild dog Lycaon pictus.
I. Diagnosing and treating causes of decline. Oryx 33: 132–142.
WORLD BANK 1983. Botswana Livestock Subsector Memorandum: Eastern Africa Projects
Department, Southern Agriculture Division, Botswana. Report No. 4665-BT.
WORTON, B.J. 1987. A review of models of home range for animal movement. Ecological
Modelling 38: 277-298.
WORTON, B.J. 1995. Using Monte Carlo simulation to evaluate kernel-based home range
estimators. Journal of Wildlife Management 59(4): 794-800.
YIRGA, G., IMAM, E., DE IONGH, H.H., LEIRS, H., KIROS, S., YOHANNES, T.G.,
TEFERI, M. & BAUER, H. 2014. Local spotted hyena abundance and community tolerance of
depredation in human-dominated landscapes in Northern Ethiopia. Mammalian Biology 79: 325–
330.
132
APPENDICES
APPENDIX I: List of the common large and medium-sized mammal species in the Northern
Tuli Game Reserve, Botswana
ORDER TUBULIDENTATA Aardvark Orycteropus afer ORDER HYRACOIDEA Rock hyrax Procavia capensis ORDER PROBOSCIDEA African elephant Loxodonta africana ORDER LAGOMORPHA Scrub hare Lepus saxatilis ORDER RODENTIA Porcupine Hysterix africaeaustralis Springhare Pedetes capensis ORDER PRIMATE Chacma baboon Papio ursinus Vervet monkey Chlorocebus pygerythrus ORDER CARNIVORA Aardwolf Proteles cristatus African civet Civettictis civetta African clawless otter Anonyx capensis African wild cat Felis lybica African wild dog Lycaon pictus Banded Mongoose Mungos mungo Bat-eared fox Otocyon megalotis Black-backed jackal Canis mesomelas Brown hyena Hyaena brunnea Cheetah Acinonyx jubatus Honey badger Mellivora capensis Large-spotted genet Genetta tigrina Leopard Panthera pardus Lion Panthera leo Selous' mongoose Paracynictis selousi Slender mongoose Galerella sanguinea Small-spotted genet Genetta genetta Spotted hyena Crocuta crocuta Striped polecat Ictonyx striatus ORDER PERISSODACTYLA Burchell’s Zebra Equus burchellii ORDER SUIFORMES Bush pig Potomachoerus porcus Warthog Phacochoerus africanus ORDER WHIPPOMORPHA Hippopotamus Hippopotamus amphibius ORDER RUMINANTIA Blue Wildebeest Connochaetes taurinus
133
Bushbuck Tragelaphus scriptus Common Duiker Sylvicapra grimmia Eland Tragelaphus oryx Giraffe Giraffa camelopardalis Impala Aepyceros melampus Klipspringer Oreotragus oreotragus Kudu Tragelaphus strepsiceros Steenbok Raphicerus campestris Waterbuck Kobus ellipsiprymnus
134
Appendix II: Photographic recording rate of mammal species during the first camera trapping
survey in the Northern Tuli Game Reserve. The index of relative abundance (RAI) is calculated
as the average number of captures per 100 trapping occasions. Species percentage (Spp. %) is
the number of capture events (n) to the total number of animal photographs.
Species n Spp. % RAI
Aardvaark Orycteropus afer 8 0.24% 0.08 Aardwoolf Proteles cristatus 3 0.09% 0.03
African Civet Civettictis civetta 2 0.06% 0.02
African Elephant Loxodonta africana 473 14.14% 4.73
African wildcat Felis lybica 5 0.15% 0.05
Banded mongoose Mungos mungo 1 0.03% 0.01
Black-backed jackal Canis mesomelas 74 2.21% 0.74
Blue wildebeest Connochaetes taurinus 71 2.12% 0.71
Burchell’s Zebra Equus burchellii 175 5.23% 1.75
Bushbuck Tragelaphus scriptus 1 0.03% 0.01
Bushpig Potomachoerus porcus 1 0.03% 0.01
Chacma Baboon Papio ursinus 133 3.97% 1.33
Cheetah Acinonyx jubatus 9 0.27% 0.09
Common duiker Sylvicapra grimmia 9 0.27% 0.09
Eland Tragelaphus oryx 211 6.31% 2.11
Giraffe Giraffa camelopardalis 250 7.47% 2.5
Honey badger Mellivora capensis 2 0.06% 0.02
Impala Aepyceros melampus 1309 39.12% 13.09
Kudu Tragelaphus strepsiceros 40 1.20% 0.4
Large-spotted genet Genetta tigrina 5 0.15% 0.05
Leopard Panthera pardus 17 0.51% 0.17
Lion Panthera leo 1 0.03% 0.01
Porcupine Hysterix africaeaustralis 6 0.18% 0.06
Scrub hare Lepus saxatilis 64 1.91% 0.64
Selous' mongoose Paracynictis selousi 3 0.09% 0.03
Small-spotted genet Genetta genetta 2 0.06% 0.02
Spotted hyena Crocuta crocuta 120 3.59% 1.2
Steenbok Raphicerus campestris 39 1.17% 0.39
135
Tree squirrel Paraxerus cepapi 5 0.15% 0.05
Vervet monkey Chlorocebus pygerythrus 13 0.39% 0.13
Warthog Phacochoerus africanus 88 2.63% 0.88
Waterbuck Kobus ellipsiprymnus 2 0.06% 0.02
136
Appendix III: Photographic recording rate of mammal species during the second camera
trapping survey in the Northern Tuli Game Reserve. The index of relative abundance (RAI) is
calculated as the average number of captures per 100 trapping occasions. Species percentage
(Spp. %) is the number of capture events (n) to the total number of animal photographs.
Species n Spp. % RAI
Aardvaark Orycteropus afer 3 0.09% 0.03 Aardwoolf Proteles cristatus 2 0.06% 0.02
African Civet Civettictis civetta 4 0.12% 0.04
Elephant Loxodonta africana 56 1.69% 0.56
African wildcat Felis lybica 1 0.03% 0.01
Banded mongoose Mungos mungo 0 0.00% 0
Bat Species unidentifiable 2 0.06% 0.02
Black-backed jackal Canis mesomelas 73 2.20% 0.73
Blue wildebeest Connochaetes taurinus 88 2.65% 0.88
Burchell’s Zebra Equus burchellii 90 2.71% 0.9
Bushbuck Tragelaphus scriptus 0 0.00% 0
Bushpig Potomachoerus porcus 1 0.03% 0.01
Chacma Baboon Papio ursinus 73 2.20% 0.73
Cheetah Acinonyx jubatus 53 1.59% 0.53
Common duiker Sylvicapra grimmia 3 0.09% 0.03
Eland Tragelaphus oryx 99 2.98% 0.99
Giraffe Giraffa camelopardalis 336 10.11% 3.36
Honey badger Mellivora capensis 1 0.03% 0.01
Impala Aepyceros melampus 213 6.41% 2.13
Kudu Tragelaphus strepsiceros 96 2.89% 0.96
Large-spotted genet Genetta tigrina 0 0.00% 0
Leopard Panthera pardus 13 0.39% 0.13
Lion Panthera leo 5 0.15% 0.05
Mouse Species unidentifiable 3 0.09% 0.03
Porcupine Hysterix africaeaustralis 8 0.24% 0.08
Scrub hare Lepus saxatilis 23 0.69% 0.23
Selous' mongoose Paracynictis selousi 0 0.00% 0
Small-spotted genet Genetta genetta 10 0.30% 0.1
137
Spotted hyena Crocuta crocuta 64 1.93% 0.64
Steenbok Raphicerus campestris 35 1.05% 0.35
Tree squirrel Paraxerus cepapi 1 0.03% 0.01
Vervet monkey Chlorocebus pygerythrus 9 0.27% 0.09
Warthog Phacochoerus africanus 23 0.69% 0.23
Waterbuck Kobus ellipsiprymnus 0 0.00% 0
138
APPENDIX IV: Cheetah Identity kit for all identified individuals within and adjacent to NOTUGRE Each cheetah was assigned a unique identity code consisting of two letters and a number. The first letter referred to the species (C = Cheetah), the second
letter, the sex (F = Female; M = Male; US = Unknown sex). The number referred to the position in the identification sequence, but is separate for males and
females. CF1 and CF5 were initially believed to be different cheetahs but were then confirmed to be the same individual and reclassified as one cheetah
CF1, thus CF5 was removed from the sequence to avoid confusion. The individual profiles are shown below for each cheetah using the best available
photographs in which the spot patterns are clearly visible. Both left hand side and right hand sides photographs of the cheetahs are included if available.
Photos provided by volunteers of the cheetah photographic survey.
FEMALE CHEETAHS RIGHT SIDE LEFT SIDE
CF1 CF2
140
RIGHT SIDE LEFT SIDE
CF3 (Mapula)
141
RIGHT SIDE LEFT SIDE CF4
CF6
142
RIGHT SIDE LEFT SIDE CF7
CF8
143
RIGHT SIDE LEFT SIDE CF9
CF10
RIGHT SIDE LEFT SIDE CF11
CF12
RIGHT SIDE LEFT SIDE CF13
No image available
CF14
No image available
RIGHT SIDE LEFT SIDE
CF15
MALE CHEETAHS RIGHT SIDE LEFT SIDE
CM1
CM2
RIGHT SIDE LEFT SIDE CM3
CM4
RIGHT SIDE LEFT SIDE CM5
CM6
RIGHT SIDE LEFT SIDE CM7
CM8
RIGHT SIDE LEFT SIDE CM9
CM10
RIGHT SIDE LEFT SIDE CM11
CM12
RIGHT SIDE LEFT SIDE CM13
No image available
CM14
RIGHT SIDE LEFT SIDE CM15
No image available
CM16
No image available
RIGHT SIDE LEFT SIDE CM17
No image available
CM18
No image available
UNKNOWN SEX CUS1 CUS2
CUS3 CUS4
CUS5 CUS6
CUS7 CUS8
CUS9 CUS10
CUS11 CUS12
CUS13
Appendix V: Description of all individual cheetahs identified during the photographic census; data period: January 2006-November 2013. The identification
number, group name, number of sightings, availability of left hand side (LHS) and right hand side (RHS) profiles are shown and estimated date of birth (year and
month), approximate age and relatedness is shown for individuals seen continuously over > 1 year. Age classes (A = Adult; C = Dependent cub/sub-adult) are
given for individuals at first sighting. The age of cheetahs are calculated for the end of 2013.
Cheetah ID
Group name
No of sightings
L H S
R H S
Age class
YOB Month Age Family relation First sighted
Last sighted
Country sighted
CF1 26 Y Y A Prior to 2005 >9 Mother of CM1, CM2; CUS10 (first litter); CM4 (second litter)
17/01/2006 08/01/2013 BOTS
CF2 Family of 2 28 Y Y A Prior to 2006 >8 Mother of CF10 18/11/2008 29/06/2012 BOTS CF3 Family of 6 155 Y Y A 2006 Jan-Feb 7.5 Mother of CF4; CF9; CF11; CM9; CM11
(first litter); CF8 (second litter) 01/10/2008 03/11/2013 BOTS
CF4 Family of 6 87 Y Y C 2011 Apr-May 2.5 Daughter of CF3; sister of CM9, CM11, CF9,CF11
26/10/2011 22/10/2013 BOTS
CF6 Female with scar 7 Y Y A Prior to 2010 >4 Mother of CF7 09/02/2012 12/01/2013 BOTS CF7 Female with scar 6 Y Y C 2011 Jul -Aug 2 Daughter of CF6 09/02/2012 12/01/2013 BOTS CF8 58 Y Y C 2013 May 7 m Daughter of CF3 24/05/2013 03/11/2013 BOTS CF9 Family of 6 79 Y Y C 2011 Apr-May 2.5 Daughter of CF3; sister of CM9, CM11,
CF4,CF11 26/10/2011 12/08/2013 BOTS
CF10 Family of 2 16 Y Y C 2010 Jun-July 3.5 Daughter of CF2 12/2/2011 05/07/2012 BOTS CF11 Family of 6 60 Y Y C 2011 Apr-May 2.5 Daughter of CF3; sister of CM9, CM11,
CF4,CF9 26/10/2011 07/12/2012 BOTS
CF12 3 Y Y A Prior to 2005 Mother of CM14 & CM15 06/08/2006 31/01/2007 BOTS CF13 1 N Y A N/A 06/01/2007 06/01/2007 RSA CF14 Moyo family 1 Y N A N/A 20/06/2013 20/06/2013 RSA CF15 Moyo family 1 Y Y SubA 2012 Daughter of CF14; sister of CM18 20/06/2013 20/06/2013 RSA CM1 Coalition of 3 78 Y Y Cub 2006 Jun-Jul 7.5 Son of CF1; brother of CM2 13/12/2006 28/10/2013 BOTS/ZIM CM2 Coalition of 3 79 Y Y Cub 2006 Jun-Jul 7.5 Son of CF1; brother of CM1 13/12/2006 28/10/2013 BOTS/ZIM CM3 Coalition of 3 75 Y Y A Prior to 2006 >7 01/05/2008 28/10/2013 BOTS/ZIM CM4 5 Y Y SubA 2009 Jun -Aug 4.5 Son of CF1 11/08/2010 15/05/2011 BOTS CM5 Coalition of 2 5 Y Y A Adult Coalition with CM6 28/06/2013 29/10/2013 BOTS CM6 Coalition of 2 6 Y Y A Adult Coalition with CM5 28/06/2013 29/10/2013 BOTS
CM7 Venetia coalition 7 Y Y A Adult Coalition with CM8 04/05/2013 09/09/2013 RSA CM8 Venetia coalition 1 Y Y A Adult Coalition with CM7 16/08/2013 16/08/2013 RSA CM9 Family of 6 74 Y Y Cub 2011 Apr-May 2.5 Son of CF3; brother of CM11, CF4, CF9,
CF11 26/10/2011 14/04/2013 BOTS
CM10 7 Y Y A 2006 Jan-Feb 2.5 possibly brother of CF3 01/10/2009 12/11/2009 BOTS CM11 Family of 6 63 Y Y Cub 2011 Apr-May Died Son of CF3; brother of CM9, CF4,
CF9, CF11 26/10/2011 16/12/2012 BOTS
CM12 Vehmbe cheetah 1 Y Y A Adult 15/09/2012 15/09/2012 RSA CM13 1 Y N A Adult 16/12/2006 16/12/2006 BOTS CM14 3 Y Y SubA 2006 Mar -May Son of CF12; brother of CM15 25/01/2007 31/01/2007 BOTS CM15 2 Y N SubA 2006 Mar -May Son of CF12; brother of CM14 25/01/2007 31/01/2007 BOTS CM16 2 N Y SubA 2006 - 2007 Brother of CM17; CUS6; CUS7 02/05/2008 29/06/2008 BOTS CM17 1 N Y SubA 2006 - 2007 Died Brother of CM16; CUS6; CUS7 02/05/2008 02/05/2008 BOTS CM18 Moyo family 1 Y N SubA 2012 Son of CF14; brother of CF15 20/06/2013 20/06/2013 RSA CUS1 2 Y N A Prior to 2007 22/03/2008 18/06/2008 BOTS CUS2 1 Y N A Prior to 2006 18/03/2006 18/03/2006 BOTS CUS3 1 N Y A Prior to 2005 17/01/2006 17/01/2006 BOTS CUS4 1 N Y SubA Prior to 2005 17/01/2006 17/01/2006 BOTS CUS5 1 Y N SubA Prior to 2005 17/01/2006 17/01/2006 BOTS CUS6 2 Y N SubA Prior to 2008 Died 02/05/2008 02/05/2008 BOTS CUS7 1 N Y SubA Prior to 2008 Died 02/05/2008 02/05/2008 BOTS CUS8 1 N Y A Prior to 2002 12/01/2003 12/01/2003 BOTS CUS9 1 N Y A 24/03/2006 24/03/2006 BOTS
CUS10 1 N Y Cub 2006 Jun - Jul Died sibling of CM1, CM2 13/12/2006 13/12/2006 BOTS CUS11 1 Y N A sibling of CUS12 21/08/2007 21/08/2007 BOTS CUS12 1 Y N A sibling of CUS11 21/08/2007 21/08/2007 BOTS CUS13 1 Y N A 01/03/2012 01/03/2012 BOTS
APPENDIX VI: Questionnaire
Date Interviewer Q.No: Coordinates S: E: Region/Village:
Section A: General Details
1. Name Anonymo us
2. Are you: owner herdman worker other 3. If you are the owner: do you own/lease/rent the land? 4. How long have you been in the area?
Section B: Socio-Economic Details
5. Year of birth? (age) 6. School level? 7. What is the main source of family income? Smallstoc Cattle Hunting Employm Crops Tourism Other……………………… 8. What role do livestock play, if it is not the main source of income?
Section C: Farm Details
9. How many persons live on cattlepost? 10. Vegetation type: 11. Can you give a percentage for each?
Open grassland and pans Sparse bush 10-35%
Medium bush 35-65%
Thick bush 65-100%
don't know
12. Have there been changes in the habitat over time?
yes no
13. Specify (how / over what time period) don't know 14. How many animals do you keep? Number Breeds Number breed Cattle Dogs Sheep Other Goats Horses Donkeys Chicken 15. Have you had major increase / decrease in livestock? How, why and over what time period? 16. What are the main problems encountered by livestock farmers? Rank importance: max 3
diseases insufficient grazing theft drought poor quality grazing snares infertility low yields vet fence losses due to predators unreliable market miscarriage
Section D: Farm Management
17. Please explain how your cattle are tended to? (e.g. people as herders, kraal, dogs) Night: Day:
18. Please explain how your goats & sheep are tended to? Night: Day:
19. If kraal Time out of kraal in morning: Time in Kraal in afternoon: Cattle: Cattle:
Goats & sheep: Goats & Sheep :
20. What is your kraal design? (stones, wooden posts, acacia bush, fencing) Cattle:
21. Distance of kraal from homestead:
22. Height: Gaps in kraal: Stems: in/out
23. Goats and sheep: 24. Distance of kraal from homestead:
25. Height: Gaps in kraal: Stems in/out
Calving/lambing
During calving / lambing, do you:
26. Bring calving animals closer to homestead? yes no
27. Check on livestock more often than before? yes no
28. Keep careful records? yes no
29. Kraal all livestock at night? yes no
30. Kraal young calves / kids during the day? yes no
31. Use a maternity / calving kraal? yes no
Other?
32. Do you have a herder with livestock? (Specify cattle/goats/sheep) yes no
33. How many do you have per number of livestock?
34. Who are they (paid workers, children)?
35. What do they do?
36. Are they effective?
37. Distance travelled to furthest grazing:
1km 1-5km 5-10km >10km
38. Do you have a dog with livestock? (Specify cattle/goats/sheep) yes no
39. What breed? 40. Size of dog: Small (cat size) Medium Large (Anatolian/German shepherd) 41. How many do you have per number of livestock? 42. Are they effective? 43. How does the dog protect
livestock: barking chasing killing predator other
Section E: Wildlife Details
45. Please detail which game species exist in the area: Species Status (absent, rare,
common, very common) Trends over past 10 years (increase, decrease, stable)
Kudu don't know Impala don't know eland don't know zebra don't know wildebeest don't know duiker don't know steenbok don't know warthog don't know hares don't know guineafowl don't know other don't know other don't know 46. Do you have explanations for any changes in the numbers? don't know
Section F: Predator Details
47. How often do you see this predator? Species
how often visual, tracks,
calls increasing, decreasing, stable
KEY
Lion never 0 Cheetah < once /year 1 Leopard once /year 2 Spotted Hyena a few times a year 3 Brown Hyena every few months 4 Wild dog once/month 5 Baboon every week 6 Jackal everyday 7 Honey Badger don't know - Other 48. Do you have explanations for any changes in the numbers? don't know
Section G: Cheetah Details
49. In the past 18 months, how many times did you see cheetahs (S)? Or have report of cheetah sightings? (R) Describe when, where, the numbers of adults and cubs and what was its activity.
Date Group Composition (number, age, sex)
Location (habitat type) Activity Time of day
50. What is the maximum number of cheetahs you have seen at the same time?
44. Does the dog live with livestock day and night?
Adults cubs (size) don't know 51. Where did you see them? Are cheetahs on your farm: permanent seasonal 52. What time of the year? 53. If you think numbers are increasing/decreasing, what are the reasons in your opinion for such as change?
don't know 54. Are cheetahs protected in yes don't know 55. Are cheetahs endangered yes don't know 56. Have cheetahs any value for you? yes no don’t know 57. Specify (traditional, totem, medicine, food, beauty, etc. OR danger, threat, problem, etc.) 58. Have your neighbours seen cheetahs? yes no don’t know
Section H: Predation and conflicts
59. Do you lose livestock to predators yes no don't know 60. Classify the predators, according to level of problem: Rank: 1: biggest problem; 10: least problem Spotted hyena Lion Honey Cheetah Wild dog Eagle Leopard Jackal Hawks Brown hyena Baboon Other… 61. If you had a problem with predators in the last 18 months, describe:
Date or Season
Animals killed or injured (no, spp, age, sex)
Predators responsible (number, spp, age, sex)
How it was identified (visual (by who), spoor by kill, killing bites, feeding style)
Time of day of incident
Location (near water, in kraal, out)
62. Are these incidents from accurate records or from memory? 63. What do you do to protect cattle from predators? 64. Kraal yes no 65. Use people as herders yes no 66. Guard animals (e.g. dogs) yes no 67. Chase away/make noise yes no 68. Use poison baits yes no 69. Other (explain) 70. What do you do to protect goats/sheep from predators? 71. Kraal yes no 72. Use people as herders yes no 73. Guard animals (e.g. dogs) yes no 74. Chase away/make noise yes no 75. Use poison baits yes no 76. Other (explain) 77. Are measures you take to protect livestock effective? yes no 78. If not, why? 79. What are the common circumstances of attacks? Day Inside kraal Herder Night Outside kraal No herder
80. Are losses to predators seasonal? yes no don't know 81. Which season? 82. Have you lost animals in the past 18 months due to other causes than predators?
Specify: number/species If no numbers: rank importance: max 3
Diseases Starvation Miscarriage Calving Theft Other Accidents Crossing vet fence Other 83. Can you give an approximate value for these losses in the last 18 months?
During your time in the area is the problem with predators: increasing decreasing stable
84. Can you give reasons why? 85. What do you do when you have a loss to a predator? (nothing, scare off predator, report to wildlife officer, kill
predator, other) 86. Did you ever have to remove predator? yes no Details: How? When? (live trap, shoot, poison) don’t know 87. Have you contacted Wildlife officer for assistance? yes no don’t know 88. Details?
Section I: Attitudes
89. What do you think about sharing the land with predators?
Benefit to farm Like them Dislike them Kill when see Other don’t know 90. Why? 91. Do you think wildlife is a national resource to be protected? yes no 92. Why? 93. Whose responsibility do you think predator/livestock conflict belongs to? don’t know
Owner Herders Government NGO's Game reserves Other
94. Do you see any solutions for the survival of predators on farmlands? Improve farm management Decrease numbers Translocate Other
Trophy hunting Compensate Tourism Contact address: Tel number: Ranking Precision Co-operative attitude Total Consistency No wrong or doubtful info /4 Predator identification Cheetah Leopard Lion Wild dog
Spotted hyena Brown hyena Domestic dog Black-backed jackal
APPENDIX VII: Predator images used in the questionnaire survey as supplementary material to assess knowledge of the common predators.
Cheetah (Acinonyx jubatus) (Photo: Eleanor Brassine) Leopard (Panthera pardus) (Photo: Eleanor Brassine)
Lion (Panthera leo) (Photo: Eleanor Brassine) African wild dog (Lycaon pictus) (Photo: Eleanor Brassine)
Spotted hyena (Crocuta crocuta) (Photo: Eleanor Brassine) Brown hyena (Hyaena brunnea) (Photo: Martin Harvey)
Domestic dog (Canis lupus familiaris) (Photo: Mathilde Brassine) Black-backed jackal (Canis mesomelas) (Photo: Eleanor Brassine)
Appendix VIII: Examples of the different kraal designs and materials utilised by livestock farmers.
Horizontal poles (split-rail fence) (Photo: Eleanor Brassine) Vertical wooden posts (Photo: Eleanor Brassine)
Acacia branches (Photo: AWF/Nakedli Maputla) Combination of wooden posts and fencing (Photo: Eleanor Brassine)
Fencing (Photo: Eleanor Brassine) Fencing with diamond mesh (Photo: Eleanor Brassine)